diff --git a/.clang_format.hook b/.clang_format.hook index 4cbc972bbd200d0dcb6d8ba404bb1286ee81736c..40d70f56cf97f7b7f18bb255dae73ab1d542f12a 100755 --- a/.clang_format.hook +++ b/.clang_format.hook @@ -1,7 +1,7 @@ #!/usr/bin/env bash set -e -readonly VERSION="3.9" +readonly VERSION="3.8" version=$(clang-format -version) diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index ede1c53a4aefc9cc144ea20f57193bc5dbb886ec..1584bc76a9dd8ddff9d05a8cb693bcbd2e09fcde 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -1,5 +1,5 @@ -- repo: https://github.com/pre-commit/mirrors-yapf.git - sha: v0.16.0 +- repo: https://github.com/PaddlePaddle/mirrors-yapf.git + sha: 0d79c0c469bab64f7229c9aca2b1186ef47f0e37 hooks: - id: yapf files: \.py$ diff --git a/.travis.yml b/.travis.yml index 52bfd5a1ba02b8ff32ef4248e00530fdd1319174..7fa098a7c5d8ecdc5d5ea38fe38c3ddae959498a 100644 --- a/.travis.yml +++ b/.travis.yml @@ -16,11 +16,12 @@ addons: - python - python-pip - python2.7-dev + - clang-format-3.8 ssh_known_hosts: 52.76.173.135 - before_install: - - sudo pip install -U virtualenv pre-commit pip - - docker pull paddlepaddle/paddle:latest + - if [[ "$JOB" == "PRE_COMMIT" ]]; then sudo ln -s /usr/bin/clang-format-3.8 /usr/bin/clang-format; fi + - sudo pip install -U virtualenv pre-commit pip + - docker pull paddlepaddle/paddle:latest script: - exit_code=0 diff --git a/README.cn.md b/README.cn.md index 8806dbfa52b152f123634f84dd1402098b017b45..72fb35ff3b239d8fa5e226f84aa09f084f593697 100644 --- a/README.cn.md +++ b/README.cn.md @@ -29,15 +29,16 @@ PaddlePaddle提供了丰富的运算单元，帮助大家以模块化的方式 点击率预估模型预判用户对一条广告点击的概率，对每次广告的点击情况做出预测，是广告技术的核心算法之一。逻谛斯克回归对大规模稀疏特征有着很好的学习能力，在点击率预估任务发展的早期一统天下。近年来，DNN 模型由于其强大的学习能力逐渐接过点击率预估任务的大旗。 -在点击率预估任务中，我们给出谷歌提出的 Wide & Deep 模型。这一模型融合了适用于学习抽象特征的DNN和适用于大规模稀疏特征的逻谛斯克回归两者的优点，可以作为一种相对成熟的模型框架使用，在工业界也有一定的应用。 +在点击率预估任务中，我们首先给出谷歌提出的 Wide & Deep 模型。这一模型融合了适用于学习抽象特征的DNN和适用于大规模稀疏特征的逻谛斯克回归两者的优点，可以作为一种相对成熟的模型框架使用，在工业界也有一定的应用。同时，我们提供基于因子分解机的深度神经网络模型，该模型融合了因子分解机和深度神经网络，分别建模输入属性之间的低阶交互和高阶交互。 -- 3.1 [Wide & deep 点击率预估模型](https://github.com/PaddlePaddle/models/tree/develop/ctr) +- 3.1 [Wide & deep 点击率预估模型](https://github.com/PaddlePaddle/models/tree/develop/ctr/README.cn.md) +- 3.2 [基于深度因子分解机的点击率预估模型](https://github.com/PaddlePaddle/models/tree/develop/deep_fm) ## 4. 文本分类 文本分类是自然语言处理领域最基础的任务之一，深度学习方法能够免除复杂的特征工程，直接使用原始文本作为输入，数据驱动地最优化分类准确率。 -在文本分类任务中，我们以情感分类任务为例，提供了基于DNN的非序列文本分类模型，以及基于CNN的序列模型供大家学习和使用（基于LSTM的模型见PaddleBook中[情感分类](https://github.com/PaddlePaddle/book/blob/develop/06.understand_sentiment/README.cn.md)一课）。 +在文本分类任务中，我们以情感分类任务为例，提供了基于DNN的非序列文本分类模型，以及基于CNN的序列模型供大家学习和使用（基于LSTM的模型见PaddleBook中[情感分类](http://www.paddlepaddle.org/docs/develop/book/06.understand_sentiment/index.cn.html)一课）。 - 4.1 [基于DNN/CNN的情感分类](https://github.com/PaddlePaddle/models/tree/develop/text_classification) - 4.2 [基于双层序列的文本分类模型](https://github.com/PaddlePaddle/models/tree/develop/nested_sequence/text_classification) @@ -46,7 +47,7 @@ PaddlePaddle提供了丰富的运算单元，帮助大家以模块化的方式 排序学习(Learning to Rank， LTR)是信息检索和搜索引擎研究的核心问题之一，通过机器学习方法学习一个分值函数对待排序的候选进行打分，再根据分值的高低确定序关系。深度神经网络可以用来建模分值函数，构成各类基于深度学习的LTR模型。 -在排序学习任务中，我们介绍基于RankLoss损失函数Pairwise排序模型和基于LambdaRank损失函数的Listwise排序模型(Pointwise学习策略见PaddleBook中[推荐系统](https://github.com/PaddlePaddle/book/blob/develop/05.recommender_system/README.cn.md)一课）。 +在排序学习任务中，我们介绍基于RankLoss损失函数Pairwise排序模型和基于LambdaRank损失函数的Listwise排序模型(Pointwise学习策略见PaddleBook中[推荐系统](http://www.paddlepaddle.org/docs/develop/book/05.recommender_system/index.cn.html)一课）。 - 5.1 [基于Pairwise和Listwise的排序学习](https://github.com/PaddlePaddle/models/tree/develop/ltr) @@ -56,7 +57,7 @@ PaddlePaddle提供了丰富的运算单元，帮助大家以模块化的方式 在结构化语义模型任务中，我们演示如何建模两个字符串之间的语义相似度。模型支持DNN(全连接前馈网络)、CNN(卷积网络)、RNN(递归神经网络)等不同的网络结构，以及分类、回归、排序等不同损失函数。本例采用最简单的文本数据作为输入，通过替换自己的训练和预测数据，便可以在真实场景中使用。 -- 6.1 [深度结构化语义模型](https://github.com/PaddlePaddle/models/tree/develop/dssm) +- 6.1 [深度结构化语义模型](https://github.com/PaddlePaddle/models/tree/develop/dssm/README.cn.md) ## 7. 命名实体识别 @@ -72,7 +73,7 @@ PaddlePaddle提供了丰富的运算单元，帮助大家以模块化的方式 在序列到序列学习任务中，我们首先以机器翻译任务为例，提供了多种改进模型供大家学习和使用。包括：不带注意力机制的序列到序列映射模型，这一模型是所有序列到序列学习模型的基础；使用Scheduled Sampling改善RNN模型在生成任务中的错误累积问题；带外部记忆机制的神经机器翻译，通过增强神经网络的记忆能力，来完成复杂的序列到序列学习任务。除机器翻译任务之外，我们也提供了一个基于深层LSTM网络生成古诗词，实现同语言生成的模型。 -- 8.1 [无注意力机制的神经机器翻译](https://github.com/PaddlePaddle/models/tree/develop/nmt_without_attention) +- 8.1 [无注意力机制的神经机器翻译](https://github.com/PaddlePaddle/models/tree/develop/nmt_without_attention/README.cn.md) - 8.2 [使用Scheduled Sampling改善翻译质量](https://github.com/PaddlePaddle/models/tree/develop/scheduled_sampling) - 8.3 [带外部记忆机制的神经机器翻译](https://github.com/PaddlePaddle/models/tree/develop/mt_with_external_memory) - 8.4 [生成古诗词](https://github.com/PaddlePaddle/models/tree/develop/generate_chinese_poetry) @@ -97,12 +98,16 @@ PaddlePaddle提供了丰富的运算单元，帮助大家以模块化的方式 图像相比文字能够提供更加生动、容易理解及更具艺术感的信息，是人们转递与交换信息的重要来源。图像分类是根据图像的语义信息对不同类别图像进行区分，是计算机视觉中重要的基础问题，也是图像检测、图像分割、物体跟踪、行为分析等其他高层视觉任务的基础，在许多领域都有着广泛的应用。如：安防领域的人脸识别和智能视频分析等，交通领域的交通场景识别，互联网领域基于内容的图像检索和相册自动归类，医学领域的图像识别等。 -在图像分类任务中，我们向大家介绍如何训练AlexNet、VGG、GoogLeNet和ResNet模型。同时提供了一个够将Caffe训练好的模型文件转换为PaddlePaddle模型文件的模型转换工具。 +在图像分类任务中，我们向大家介绍如何训练AlexNet、VGG、GoogLeNet、ResNet、Inception-v4、Inception-Resnet-V2和Xception模型。同时提供了能够将Caffe或TensorFlow训练好的模型文件转换为PaddlePaddle模型文件的模型转换工具。 - 11.1 [将Caffe模型文件转换为PaddlePaddle模型文件](https://github.com/PaddlePaddle/models/tree/develop/image_classification/caffe2paddle) -- 11.2 [AlexNet](https://github.com/PaddlePaddle/models/tree/develop/image_classification) -- 11.3 [VGG](https://github.com/PaddlePaddle/models/tree/develop/image_classification) -- 11.4 [Residual Network](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 11.2 [将TensorFlow模型文件转换为PaddlePaddle模型文件](https://github.com/PaddlePaddle/models/tree/develop/image_classification/tf2paddle) +- 11.3 [AlexNet](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 11.4 [VGG](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 11.5 [Residual Network](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 11.6 [Inception-v4](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 11.7 [Inception-Resnet-V2](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 11.8 [Xception](https://github.com/PaddlePaddle/models/tree/develop/image_classification) ## 12. 目标检测 @@ -110,7 +115,7 @@ PaddlePaddle提供了丰富的运算单元，帮助大家以模块化的方式 在目标检测任务中，我们介绍利用SSD方法完成目标检测。SSD全称：Single Shot MultiBox Detector，是目标检测领域较新且效果较好的检测算法之一，具有检测速度快且检测精度高的特点。 -- 12.1 [Single Shot MultiBox Detector](https://github.com/PaddlePaddle/models/tree/develop/ssd) +- 12.1 [Single Shot MultiBox Detector](https://github.com/PaddlePaddle/models/tree/develop/ssd/README.cn.md) ## 13. 场景文字识别 @@ -126,7 +131,6 @@ PaddlePaddle提供了丰富的运算单元，帮助大家以模块化的方式 在语音识别任务中，我们提供了基于 DeepSpeech2 模型的完整流水线，包括：特征提取、数据增强、模型训练、语言模型、解码模块等，并提供一个训练好的模型和体验实例，大家能够使用自己的声音来体验语音识别的乐趣。 -- 14.1 [语音识别: DeepSpeech2](https://github.com/PaddlePaddle/models/tree/develop/deep_speech_2) - +14.1 [语音识别: DeepSpeech2](https://github.com/PaddlePaddle/DeepSpeech) 本教程由[PaddlePaddle](https://github.com/PaddlePaddle/Paddle)创作，采用[Apache-2.0](LICENSE) 许可协议进行许可。 diff --git a/README.md b/README.md index 3b2da82aeebb5fc0535144940965598c8d9dfc02..1920b307253c9cc2f4547df1886124f85bfdabb2 100644 --- a/README.md +++ b/README.md @@ -25,15 +25,16 @@ The language model is important in the field of natural language processing. In ## 3. Click-Through Rate prediction The click-through rate model predicts the probability that a user will click on an ad. This is widely used for advertising technology. Logistic Regression has a good learning performance for large-scale sparse features in the early stages of the development of click-through rate prediction. In recent years, DNN model because of its strong learning ability to gradually take the banner rate of the task of the banner. -In the example of click-through rate estimates, we give the Google's Wide & Deep model. This model combines the advantages of DNN and the applicable logistic regression model for DNN and large-scale sparse features. +In the example of click-through rate estimates, we first give the Google's Wide & Deep model. This model combines the advantages of DNN and the applicable logistic regression model for DNN and large-scale sparse features. Then we provide the deep factorization machine for click-through rate prediction. The deep factorization machine combines the factorization machine and deep neural networks to model both low order and high order interactions of input features. - 3.1 [Click-Through Rate Model](https://github.com/PaddlePaddle/models/tree/develop/ctr) +- 3.2 [Deep Factorization Machine for Click-Through Rate prediction](https://github.com/PaddlePaddle/models/tree/develop/deep_fm) ## 4. Text classification Text classification is one of the most basic tasks in natural language processing. The deep learning method can eliminate the complex feature engineering, and use the original text as input to optimize the classification accuracy. -For text classification, we provide a non-sequential text classification model based on DNN and CNN. (For LSTM-based model, please refer to PaddleBook [Sentiment Analysis](https://github.com/PaddlePaddle/book/blob/develop/06.understand_sentiment/README.cn.md)). +For text classification, we provide a non-sequential text classification model based on DNN and CNN. (For LSTM-based model, please refer to PaddleBook [Sentiment Analysis](http://www.paddlepaddle.org/docs/develop/book/06.understand_sentiment/index.html)). - 4.1 [Sentiment analysis based on DNN / CNN](https://github.com/PaddlePaddle/models/tree/develop/text_classification) @@ -42,7 +43,7 @@ For text classification, we provide a non-sequential text classification model b Learning to rank (LTR) is one of the core problems in information retrieval and search engine research. Training data is used by a learning algorithm to produce a ranking model which computes the relevance of documents for actual queries. The depth neural network can be used to model the fractional function to form various LTR models based on depth learning. -The algorithms for learning to rank are usually categorized into three groups by their input representation and the loss function. These are pointwise, pairwise and listwise approaches. Here we demonstrate RankLoss loss function method (pairwise approach), and LambdaRank loss function method (listwise approach). (For Pointwise approaches, please refer to [Recommended System](https://github.com/PaddlePaddle/book/blob/develop/05.recommender_system/README.cn.md)). +The algorithms for learning to rank are usually categorized into three groups by their input representation and the loss function. These are pointwise, pairwise and listwise approaches. Here we demonstrate RankLoss loss function method (pairwise approach), and LambdaRank loss function method (listwise approach). (For Pointwise approaches, please refer to [Recommended System](http://www.paddlepaddle.org/docs/develop/book/05.recommender_system/index.html)). - 5.1 [Learning to rank based on Pairwise and Listwise approches](https://github.com/PaddlePaddle/models/tree/develop/ltr) @@ -71,11 +72,15 @@ As an example for sequence-to-sequence learning, we take the machine translation ## 9. Image classification -For the example of image classification, we show you how to train AlexNet, VGG, GoogLeNet and ResNet models in PaddlePaddle. It also provides a model conversion tool that converts Caffe trained model files into PaddlePaddle model files. +For the example of image classification, we show you how to train AlexNet, VGG, GoogLeNet, ResNet, Inception-v4, Inception-Resnet-V2 and Xception models in PaddlePaddle. It also provides model conversion tools that convert Caffe or TensorFlow trained model files into PaddlePaddle model files. - 9.1 [convert Caffe model file to PaddlePaddle model file](https://github.com/PaddlePaddle/models/tree/develop/image_classification/caffe2paddle) -- 9.2 [AlexNet](https://github.com/PaddlePaddle/models/tree/develop/image_classification) -- 9.3 [VGG](https://github.com/PaddlePaddle/models/tree/develop/image_classification) -- 9.4 [Residual Network](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 9.2 [convert TensorFlow model file to PaddlePaddle model file](https://github.com/PaddlePaddle/models/tree/develop/image_classification/tf2paddle) +- 9.3 [AlexNet](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 9.4 [VGG](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 9.5 [Residual Network](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 9.6 [Inception-v4](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 9.7 [Inception-Resnet-V2](https://github.com/PaddlePaddle/models/tree/develop/image_classification) +- 9.8 [Xception](https://github.com/PaddlePaddle/models/tree/develop/image_classification) This tutorial is contributed by [PaddlePaddle](https://github.com/PaddlePaddle/Paddle) and licensed under the [Apache-2.0 license](LICENSE). diff --git a/conv_seq2seq/README.md b/conv_seq2seq/README.md index 75ea8770266cc277843608de8320d74d54d1e8e4..920c664562ef160699dff7b391aa3f4ad8608387 100644 --- a/conv_seq2seq/README.md +++ b/conv_seq2seq/README.md @@ -4,55 +4,63 @@ This model implements the work in the following paper: Jonas Gehring, Micheal Auli, David Grangier, et al. Convolutional Sequence to Sequence Learning. Association for Computational Linguistics (ACL), 2017 # Data Preparation +- The data used in this tutorial can be downloaded by runing: -- In this tutorial, each line in a data file contains one sample and each sample consists of a source sentence and a target sentence. And the two sentences are seperated by '\t'. So, to use your own data, it should be organized as follows: + ```bash + sh download.sh + ``` - ``` - \t - ``` +- Each line in the data file contains one sample and each sample consists of a source sentence and a target sentence. And the two sentences are seperated by '\t'. So, to use your own data, it should be organized as follows: + + ``` + \t + ``` # Training a Model - Modify the following script if needed and then run: - ```bash - python train.py \ - --train_data_path ./data/train_data \ - --test_data_path ./data/test_data \ + ```bash + python train.py \ + --train_data_path ./data/train \ + --test_data_path ./data/test \ --src_dict_path ./data/src_dict \ --trg_dict_path ./data/trg_dict \ --enc_blocks "[(256, 3)] * 5" \ --dec_blocks "[(256, 3)] * 3" \ --emb_size 256 \ --pos_size 200 \ - --drop_rate 0.1 \ + --drop_rate 0.2 \ + --use_bn False \ --use_gpu False \ --trainer_count 1 \ --batch_size 32 \ --num_passes 20 \ >train.log 2>&1 - ``` + ``` # Inferring by a Trained Model - Infer by a trained model by running: - ```bash - python infer.py \ - --infer_data_path ./data/infer_data \ + ```bash + python infer.py \ + --infer_data_path ./data/dev \ --src_dict_path ./data/src_dict \ --trg_dict_path ./data/trg_dict \ --enc_blocks "[(256, 3)] * 5" \ --dec_blocks "[(256, 3)] * 3" \ --emb_size 256 \ --pos_size 200 \ - --drop_rate 0.1 \ + --drop_rate 0.2 \ + --use_bn False \ --use_gpu False \ --trainer_count 1 \ --max_len 100 \ + --batch_size 256 \ --beam_size 1 \ + --is_show_attention False \ --model_path ./params.pass-0.tar.gz \ 1>infer_result 2>infer.log ``` # Notes - -Currently, beam search will forward the encoder multiple times when predicting each target word, which requires extra computations. And we will fix it later. +Since PaddlePaddle of current version doesn't support weight normalization, we use batch normalization instead to confirm convergence when the network is deep. diff --git a/conv_seq2seq/beamsearch.py b/conv_seq2seq/beamsearch.py index 45656e809b4000af4b502bda157a4864ad3fff3a..dd8562f018c803d4f0d7bbba4a2a006ece904851 100644 --- a/conv_seq2seq/beamsearch.py +++ b/conv_seq2seq/beamsearch.py @@ -2,13 +2,15 @@ import sys import time +import math import numpy as np +import reader + class BeamSearch(object): """ Generate sequence by beam search - NOTE: this class only implements generating one sentence at a time. """ def __init__(self, @@ -16,44 +18,44 @@ class BeamSearch(object): trg_dict, pos_size, padding_num, + batch_size=1, beam_size=1, max_len=100): self.inferer = inferer self.trg_dict = trg_dict + self.reverse_trg_dict = reader.get_reverse_dict(trg_dict) self.word_padding = trg_dict.__len__() self.pos_size = pos_size self.pos_padding = pos_size self.padding_num = padding_num self.win_len = padding_num + 1 self.max_len = max_len + self.batch_size = batch_size self.beam_size = beam_size - def get_beam_input(self, pre_beam_list, infer_data): + def get_beam_input(self, batch, sample_list): """ Get input for generation at the current iteration. """ beam_input = [] - if len(pre_beam_list) == 0: - cur_trg = [self.word_padding - ] * self.padding_num + [self.trg_dict['']] - cur_trg_pos = [self.pos_padding] * self.padding_num + [0] - beam_input.append(infer_data + [cur_trg] + [cur_trg_pos]) - else: - for seq in pre_beam_list: - if len(seq) < self.win_len: + for sample_id in sample_list: + for path in self.candidate_path[sample_id]: + if len(path['seq']) < self.win_len: cur_trg = [self.word_padding] * ( - self.win_len - len(seq) - 1 - ) + [self.trg_dict['']] + seq + self.win_len - len(path['seq']) - 1 + ) + [self.trg_dict['']] + path['seq'] cur_trg_pos = [self.pos_padding] * ( - self.win_len - len(seq) - 1) + [0] + range(1, - len(seq) + 1) + self.win_len - len(path['seq']) - 1) + [0] + range( + 1, len(path['seq']) + 1) else: - cur_trg = seq[-self.win_len:] + cur_trg = path['seq'][-self.win_len:] cur_trg_pos = range( - len(seq) + 1 - self.win_len, len(seq) + 1) + len(path['seq']) + 1 - self.win_len, + len(path['seq']) + 1) + + beam_input.append(batch[sample_id] + [cur_trg] + [cur_trg_pos]) - beam_input.append(infer_data + [cur_trg] + [cur_trg_pos]) return beam_input def get_prob(self, beam_input): @@ -64,100 +66,132 @@ class BeamSearch(object): prob = self.inferer.infer(beam_input, field='value')[row_list, :] return prob - def get_candidate(self, pre_beam_list, pre_beam_score, prob): + def _top_k(self, prob, k): """ - Get top beam_size tokens and their scores for each beam. + Get indices of the words with k highest probablities. """ - if prob.ndim == 1: - candidate_id = prob.argsort()[-self.beam_size:][::-1] - candidate_log_prob = np.log(prob[candidate_id]) - else: - candidate_id = prob.argsort()[:, -self.beam_size:][:, ::-1] - candidate_log_prob = np.zeros_like(candidate_id).astype('float32') - for j in range(len(pre_beam_list)): - candidate_log_prob[j, :] = np.log(prob[j, candidate_id[j, :]]) - - if pre_beam_score.size > 0: - candidate_score = candidate_log_prob + pre_beam_score.reshape( - (pre_beam_score.size, 1)) - else: - candidate_score = candidate_log_prob - - return candidate_id, candidate_score - - def prune(self, candidate_id, candidate_score, pre_beam_list, - completed_seq_list, completed_seq_score, completed_seq_min_score): + return prob.argsort()[-k:][::-1] + + def beam_expand(self, prob, sample_list): """ - Pruning process of the beam search. During the process, beam_size most possible sequences - are selected for the beam in the next iteration. Besides, their scores and the minimum score - of the completed sequences are updated. + In every iteration step, the model predicts the possible next words. + For each input sentence, the top beam_size words are selected as candidates. """ - candidate_id = candidate_id.flatten() - candidate_score = candidate_score.flatten() - - topk_idx = candidate_score.argsort()[-self.beam_size:][::-1].tolist() - topk_seq_idx = [idx / self.beam_size for idx in topk_idx] - - next_beam = [] - beam_score = [] - for j in range(len(topk_idx)): - if candidate_id[topk_idx[j]] == self.trg_dict['']: - if len( - completed_seq_list - ) < self.beam_size or completed_seq_min_score <= candidate_score[ - topk_idx[j]]: - completed_seq_list.append(pre_beam_list[topk_seq_idx[j]]) - completed_seq_score.append(candidate_score[topk_idx[j]]) - - if completed_seq_min_score is None or ( - completed_seq_min_score >= - candidate_score[topk_idx[j]] and - len(completed_seq_list) < self.beam_size): - completed_seq_min_score = candidate_score[topk_idx[j]] - else: - seq = pre_beam_list[topk_seq_idx[ - j]] + [candidate_id[topk_idx[j]]] - score = candidate_score[topk_idx[j]] - next_beam.append(seq) - beam_score.append(score) - - beam_score = np.array(beam_score) - return next_beam, beam_score, completed_seq_min_score - - def search_one_sample(self, infer_data): + top_words = np.apply_along_axis(self._top_k, 1, prob, self.beam_size) + + candidate_words = [[]] * len(self.candidate_path) + idx = 0 + + for sample_id in sample_list: + for seq_id, path in enumerate(self.candidate_path[sample_id]): + for w in top_words[idx, :]: + score = path['score'] + math.log(prob[idx, w]) + candidate_words[sample_id] = candidate_words[sample_id] + [{ + 'word': w, + 'score': score, + 'seq_id': seq_id + }] + idx = idx + 1 + + return candidate_words + + def beam_shrink(self, candidate_words, sample_list): """ - Beam search process for one sample. + Pruning process of the beam search. During the process, beam_size most post possible + sequences are selected for the beam in the next generation. """ - completed_seq_list = [] - completed_seq_score = [] - completed_seq_min_score = None - uncompleted_seq_list = [[]] - uncompleted_seq_score = np.zeros(0) + new_path = [[]] * len(self.candidate_path) + + for sample_id in sample_list: + beam_words = sorted( + candidate_words[sample_id], + key=lambda x: x['score'], + reverse=True)[:self.beam_size] + + complete_seq_min_score = None + complete_path_num = len(self.complete_path[sample_id]) + + if complete_path_num > 0: + complete_seq_min_score = min(self.complete_path[sample_id], + key=lambda x: x['score'])['score'] + if complete_path_num >= self.beam_size: + beam_words_max_score = beam_words[0]['score'] + if beam_words_max_score < complete_seq_min_score: + continue + + for w in beam_words: + + if w['word'] == self.trg_dict['']: + if complete_path_num < self.beam_size or complete_seq_min_score <= w[ + 'score']: + + seq = self.candidate_path[sample_id][w['seq_id']]['seq'] + self.complete_path[sample_id] = self.complete_path[ + sample_id] + [{ + 'seq': seq, + 'score': w['score'] + }] + + if complete_seq_min_score is None or complete_seq_min_score > w[ + 'score']: + complete_seq_min_score = w['score'] + else: + seq = self.candidate_path[sample_id][w['seq_id']]['seq'] + [ + w['word'] + ] + new_path[sample_id] = new_path[sample_id] + [{ + 'seq': seq, + 'score': w['score'] + }] - for i in xrange(self.max_len): - beam_input = self.get_beam_input(uncompleted_seq_list, infer_data) + return new_path - prob = self.get_prob(beam_input) + def search_one_batch(self, batch): + """ + Perform beam search on one mini-batch. + """ + real_size = len(batch) + self.candidate_path = [[{'seq': [], 'score': 0.}]] * real_size + self.complete_path = [[]] * real_size + sample_list = range(real_size) - candidate_id, candidate_score = self.get_candidate( - uncompleted_seq_list, uncompleted_seq_score, prob) + for i in xrange(self.max_len): + beam_input = self.get_beam_input(batch, sample_list) + prob = self.get_prob(beam_input) - uncompleted_seq_list, uncompleted_seq_score, completed_seq_min_score = self.prune( - candidate_id, candidate_score, uncompleted_seq_list, - completed_seq_list, completed_seq_score, - completed_seq_min_score) + candidate_words = self.beam_expand(prob, sample_list) + new_path = self.beam_shrink(candidate_words, sample_list) + self.candidate_path = new_path + sample_list = [ + sample_id for sample_id in sample_list + if len(new_path[sample_id]) > 0 + ] - if len(uncompleted_seq_list) == 0: + if len(sample_list) == 0: break - if len(completed_seq_list) >= self.beam_size: - seq_max_score = uncompleted_seq_score.max() - if seq_max_score < completed_seq_min_score: - uncompleted_seq_list = [] - break - - final_seq_list = completed_seq_list + uncompleted_seq_list - final_score = np.concatenate( - (np.array(completed_seq_score), uncompleted_seq_score)) - max_id = final_score.argmax() - top_seq = final_seq_list[max_id] - return top_seq + + final_path = [] + for i in xrange(real_size): + top_path = sorted( + self.complete_path[i] + self.candidate_path[i], + key=lambda x: x['score'], + reverse=True)[:self.beam_size] + final_path.append(top_path) + return final_path + + def search(self, infer_data): + """ + Perform beam search on all data. + """ + + def _to_sentence(seq): + raw_sentence = [self.reverse_trg_dict[id] for id in seq] + sentence = " ".join(raw_sentence) + return sentence + + for pos in xrange(0, len(infer_data), self.batch_size): + batch = infer_data[pos:min(pos + self.batch_size, len(infer_data))] + self.final_path = self.search_one_batch(batch) + for top_path in self.final_path: + print _to_sentence(top_path[0]['seq']) + sys.stdout.flush() diff --git a/conv_seq2seq/download.sh b/conv_seq2seq/download.sh new file mode 100644 index 0000000000000000000000000000000000000000..b1a924d25b1a10ade9f4be8b504933d1efa01905 --- /dev/null +++ b/conv_seq2seq/download.sh @@ -0,0 +1,22 @@ +#!/usr/bin/env bash + +CUR_PATH=`pwd` +git clone https://github.com/moses-smt/mosesdecoder.git +git clone https://github.com/rizar/actor-critic-public + +export MOSES=`pwd`/mosesdecoder +export LVSR=`pwd`/actor-critic-public + +cd actor-critic-public/exp/ted +sh create_dataset.sh + +cd $CUR_PATH +mkdir data +cp actor-critic-public/exp/ted/prep/*-* data/ +cp actor-critic-public/exp/ted/vocab.* data/ + +cd data +python ../preprocess.py + +cd .. +rm -rf actor-critic-public mosesdecoder diff --git a/conv_seq2seq/infer.py b/conv_seq2seq/infer.py index eb46df5549f11217fa2881b8657b2b6add4bd7ce..c804a84e71ffe920b72064cb05461d72c444ac73 100644 --- a/conv_seq2seq/infer.py +++ b/conv_seq2seq/infer.py @@ -36,7 +36,7 @@ def parse_args(): parser.add_argument( '--emb_size', type=int, - default=512, + default=256, help='Dimension of word embedding. (default: %(default)s)') parser.add_argument( '--pos_size', @@ -48,6 +48,11 @@ def parse_args(): type=float, default=0., help='Dropout rate. (default: %(default)s)') + parser.add_argument( + "--use_bn", + default=False, + type=distutils.util.strtobool, + help="Use batch normalization or not. (default: %(default)s)") parser.add_argument( "--use_gpu", default=False, @@ -64,36 +69,43 @@ def parse_args(): default=100, help="The maximum length of the sentence to be generated. (default: %(default)s)" ) + parser.add_argument( + "--batch_size", + default=1, + type=int, + help="Size of a mini-batch. (default: %(default)s)") parser.add_argument( "--beam_size", default=1, type=int, - help="The width of beam expasion. (default: %(default)s)") + help="The width of beam expansion. (default: %(default)s)") parser.add_argument( "--model_path", type=str, required=True, help="The path of trained model. (default: %(default)s)") + parser.add_argument( + "--is_show_attention", + default=False, + type=distutils.util.strtobool, + help="Whether to show attention weight or not. (default: %(default)s)") return parser.parse_args() -def to_sentence(seq, dictionary): - raw_sentence = [dictionary[id] for id in seq] - sentence = " ".join(raw_sentence) - return sentence - - def infer(infer_data_path, src_dict_path, trg_dict_path, model_path, enc_conv_blocks, dec_conv_blocks, - emb_dim=512, + emb_dim=256, pos_size=200, drop_rate=0., + use_bn=False, max_len=100, - beam_size=1): + batch_size=1, + beam_size=1, + is_show_attention=False): """ Inference. @@ -120,10 +132,14 @@ def infer(infer_data_path, :type pos_size: int :param drop_rate: Dropout rate. :type drop_rate: float + :param use_bn: Whether to use batch normalization or not. False is the default value. + :type use_bn: bool :param max_len: The maximum length of the sentence to be generated. :type max_len: int :param beam_size: The width of beam expansion. :type beam_size: int + :param is_show_attention: Whether to show attention weight or not. False is the default value. + :type is_show_attention: bool """ # load dict src_dict = reader.load_dict(src_dict_path) @@ -131,7 +147,7 @@ def infer(infer_data_path, src_dict_size = src_dict.__len__() trg_dict_size = trg_dict.__len__() - prob = conv_seq2seq( + prob, weight = conv_seq2seq( src_dict_size=src_dict_size, trg_dict_size=trg_dict_size, pos_size=pos_size, @@ -139,6 +155,7 @@ def infer(infer_data_path, enc_conv_blocks=enc_conv_blocks, dec_conv_blocks=dec_conv_blocks, drop_rate=drop_rate, + with_bn=use_bn, is_infer=True) # load parameters @@ -153,6 +170,26 @@ def infer(infer_data_path, pos_size=pos_size, padding_num=padding_num) + if is_show_attention: + attention_inferer = paddle.inference.Inference( + output_layer=weight, parameters=parameters) + for i, data in enumerate(infer_reader()): + src_len = len(data[0]) + trg_len = len(data[2]) + attention_weight = attention_inferer.infer( + [data], field='value', flatten_result=False) + attention_weight = [ + weight.reshape((trg_len, src_len)) + for weight in attention_weight + ] + print attention_weight + break + return + + infer_data = [] + for i, raw_data in enumerate(infer_reader()): + infer_data.append([raw_data[0], raw_data[1]]) + inferer = paddle.inference.Inference( output_layer=prob, parameters=parameters) @@ -162,15 +199,10 @@ def infer(infer_data_path, pos_size=pos_size, padding_num=padding_num, max_len=max_len, + batch_size=batch_size, beam_size=beam_size) - reverse_trg_dict = reader.get_reverse_dict(trg_dict) - for i, raw_data in enumerate(infer_reader()): - infer_data = [raw_data[0], raw_data[1]] - result = searcher.search_one_sample(infer_data) - sentence = to_sentence(result, reverse_trg_dict) - print sentence - sys.stdout.flush() + searcher.search(infer_data) return @@ -179,6 +211,8 @@ def main(): enc_conv_blocks = eval(args.enc_blocks) dec_conv_blocks = eval(args.dec_blocks) + sys.setrecursionlimit(10000) + paddle.init(use_gpu=args.use_gpu, trainer_count=args.trainer_count) infer( @@ -191,8 +225,11 @@ def main(): emb_dim=args.emb_size, pos_size=args.pos_size, drop_rate=args.drop_rate, + use_bn=args.use_bn, max_len=args.max_len, - beam_size=args.beam_size) + batch_size=args.batch_size, + beam_size=args.beam_size, + is_show_attention=args.is_show_attention) if __name__ == '__main__': diff --git a/conv_seq2seq/model.py b/conv_seq2seq/model.py index 85f23862ce53871edc37f2c0a617f0130798a66b..c31238f83172fdc3d6240095279d1c953ab272ae 100644 --- a/conv_seq2seq/model.py +++ b/conv_seq2seq/model.py @@ -12,7 +12,8 @@ def gated_conv_with_batchnorm(input, context_len, context_start=None, learning_rate=1.0, - drop_rate=0.): + drop_rate=0., + with_bn=False): """ Definition of the convolution block. @@ -30,6 +31,9 @@ def gated_conv_with_batchnorm(input, :type learning_rate: float :param drop_rate: Dropout rate. :type drop_rate: float + :param with_bn: Whether to use batch normalization or not. False is the default + value. + :type with_bn: bool :return: The output of the convolution block. :rtype: LayerOutput """ @@ -50,18 +54,18 @@ def gated_conv_with_batchnorm(input, learning_rate=learning_rate), bias_attr=False) - batch_norm_conv = paddle.layer.batch_norm( - input=raw_conv, - act=paddle.activation.Linear(), - param_attr=paddle.attr.Param(learning_rate=learning_rate)) + if with_bn: + raw_conv = paddle.layer.batch_norm( + input=raw_conv, + act=paddle.activation.Linear(), + param_attr=paddle.attr.Param(learning_rate=learning_rate)) with paddle.layer.mixed(size=size) as conv: - conv += paddle.layer.identity_projection( - batch_norm_conv, size=size, offset=0) + conv += paddle.layer.identity_projection(raw_conv, size=size, offset=0) with paddle.layer.mixed(size=size, act=paddle.activation.Sigmoid()) as gate: gate += paddle.layer.identity_projection( - batch_norm_conv, size=size, offset=size) + raw_conv, size=size, offset=size) with paddle.layer.mixed(size=size) as gated_conv: gated_conv += paddle.layer.dotmul_operator(conv, gate) @@ -73,7 +77,8 @@ def encoder(token_emb, pos_emb, conv_blocks=[(256, 3)] * 5, num_attention=3, - drop_rate=0.1): + drop_rate=0., + with_bn=False): """ Definition of the encoder. @@ -89,6 +94,9 @@ def encoder(token_emb, :type num_attention: int :param drop_rate: Dropout rate. :type drop_rate: float + :param with_bn: Whether to use batch normalization or not. False is the default + value. + :type with_bn: bool :return: The input token encoding. :rtype: LayerOutput """ @@ -124,7 +132,8 @@ def encoder(token_emb, size=size, context_len=context_len, learning_rate=1.0 / (2.0 * num_attention), - drop_rate=drop_rate) + drop_rate=drop_rate, + with_bn=with_bn) with paddle.layer.mixed(size=size) as block_output: block_output += paddle.layer.identity_projection(residual) @@ -165,7 +174,7 @@ def attention(decoder_state, cur_embedding, encoded_vec, encoded_sum): :type encoded_vec: LayerOutput :param encoded_sum: The sum of the source token's encoding and embedding. :type encoded_sum: LayerOutput - :return: A context vector. + :return: A context vector and the attention weight. :rtype: LayerOutput """ residual = decoder_state @@ -182,31 +191,29 @@ def attention(decoder_state, cur_embedding, encoded_vec, encoded_sum): expanded = paddle.layer.expand(input=state_summary, expand_as=encoded_vec) - m = paddle.layer.linear_comb(weights=expanded, vectors=encoded_vec) + m = paddle.layer.dot_prod(input1=expanded, input2=encoded_vec) - attention_weight = paddle.layer.fc( - input=m, - size=1, - act=paddle.activation.SequenceSoftmax(), - bias_attr=False) + attention_weight = paddle.layer.fc(input=m, + size=1, + act=paddle.activation.SequenceSoftmax(), + bias_attr=False) scaled = paddle.layer.scaling(weight=attention_weight, input=encoded_sum) attended = paddle.layer.pooling( input=scaled, pooling_type=paddle.pooling.Sum()) - attended_proj = paddle.layer.fc( - input=attended, - size=state_size, - act=paddle.activation.Linear(), - bias_attr=True) + attended_proj = paddle.layer.fc(input=attended, + size=state_size, + act=paddle.activation.Linear(), + bias_attr=True) attention_result = paddle.layer.addto(input=[attended_proj, residual]) # halve the variance of the sum attention_result = paddle.layer.slope_intercept( input=attention_result, slope=math.sqrt(0.5)) - return attention_result + return attention_result, attention_weight def decoder(token_emb, @@ -215,7 +222,8 @@ def decoder(token_emb, encoded_sum, dict_size, conv_blocks=[(256, 3)] * 3, - drop_rate=0.1): + drop_rate=0., + with_bn=False): """ Definition of the decoder. @@ -235,7 +243,10 @@ def decoder(token_emb, :type conv_blocks: list of tuple :param drop_rate: Dropout rate. :type drop_rate: float - :return: The probability of the predicted token. + :param with_bn: Whether to use batch normalization or not. False is the default + value. + :type with_bn: bool + :return: The probability of the predicted token and the attention weights. :rtype: LayerOutput """ @@ -261,22 +272,23 @@ def decoder(token_emb, initial_std=math.sqrt((1.0 - drop_rate) / embedding.size)), bias_attr=True, ) + weight = [] for (size, context_len) in conv_blocks: if block_input.size == size: residual = block_input else: - residual = paddle.layer.fc( - input=block_input, - size=size, - act=paddle.activation.Linear(), - bias_attr=True) + residual = paddle.layer.fc(input=block_input, + size=size, + act=paddle.activation.Linear(), + bias_attr=True) decoder_state = gated_conv_with_batchnorm( input=block_input, size=size, context_len=context_len, context_start=0, - drop_rate=drop_rate) + drop_rate=drop_rate, + with_bn=with_bn) group_inputs = [ decoder_state, @@ -285,8 +297,9 @@ def decoder(token_emb, paddle.layer.StaticInput(input=encoded_sum), ] - conditional = paddle.layer.recurrent_group( + conditional, attention_weight = paddle.layer.recurrent_group( step=attention_step, input=group_inputs) + weight.append(attention_weight) block_output = paddle.layer.addto(input=[conditional, residual]) @@ -312,7 +325,7 @@ def decoder(token_emb, initial_std=math.sqrt((1.0 - drop_rate) / block_output.size)), bias_attr=True) - return decoder_out + return decoder_out, weight def conv_seq2seq(src_dict_size, @@ -321,7 +334,8 @@ def conv_seq2seq(src_dict_size, emb_dim, enc_conv_blocks=[(256, 3)] * 5, dec_conv_blocks=[(256, 3)] * 3, - drop_rate=0.1, + drop_rate=0., + with_bn=False, is_infer=False): """ Definition of convolutional sequence-to-sequence network. @@ -345,6 +359,8 @@ def conv_seq2seq(src_dict_size, :type dec_conv_blocks: list of tuple :param drop_rate: Dropout rate. :type drop_rate: float + :param with_bn: Whether to use batch normalization or not. False is the default value. + :type with_bn: bool :param is_infer: Whether infer or not. :type is_infer: bool :return: Cost or output layer. @@ -362,12 +378,14 @@ def conv_seq2seq(src_dict_size, input=src, size=emb_dim, name='src_word_emb', - param_attr=paddle.attr.Param(initial_mean=0., initial_std=0.1)) + param_attr=paddle.attr.Param( + initial_mean=0., initial_std=0.1)) src_pos_emb = paddle.layer.embedding( input=src_pos, size=emb_dim, name='src_pos_emb', - param_attr=paddle.attr.Param(initial_mean=0., initial_std=0.1)) + param_attr=paddle.attr.Param( + initial_mean=0., initial_std=0.1)) num_attention = len(dec_conv_blocks) encoded_vec, encoded_sum = encoder( @@ -375,7 +393,8 @@ def conv_seq2seq(src_dict_size, pos_emb=src_pos_emb, conv_blocks=enc_conv_blocks, num_attention=num_attention, - drop_rate=drop_rate) + drop_rate=drop_rate, + with_bn=with_bn) trg = paddle.layer.data( name='trg_word', @@ -390,24 +409,27 @@ def conv_seq2seq(src_dict_size, input=trg, size=emb_dim, name='trg_word_emb', - param_attr=paddle.attr.Param(initial_mean=0., initial_std=0.1)) + param_attr=paddle.attr.Param( + initial_mean=0., initial_std=0.1)) trg_pos_emb = paddle.layer.embedding( input=trg_pos, size=emb_dim, name='trg_pos_emb', - param_attr=paddle.attr.Param(initial_mean=0., initial_std=0.1)) + param_attr=paddle.attr.Param( + initial_mean=0., initial_std=0.1)) - decoder_out = decoder( + decoder_out, weight = decoder( token_emb=trg_emb, pos_emb=trg_pos_emb, encoded_vec=encoded_vec, encoded_sum=encoded_sum, dict_size=trg_dict_size, conv_blocks=dec_conv_blocks, - drop_rate=drop_rate) + drop_rate=drop_rate, + with_bn=with_bn) if is_infer: - return decoder_out + return decoder_out, weight trg_next_word = paddle.layer.data( name='trg_next_word', diff --git a/conv_seq2seq/preprocess.py b/conv_seq2seq/preprocess.py new file mode 100644 index 0000000000000000000000000000000000000000..1d5c7cdd7b5cc91e28854fa0bbeeffc9dcbe4e5c --- /dev/null +++ b/conv_seq2seq/preprocess.py @@ -0,0 +1,30 @@ +#coding=utf-8 + +import cPickle + + +def concat_file(file1, file2, dst_file): + with open(dst_file, 'w') as dst: + with open(file1) as f1: + with open(file2) as f2: + for i, (line1, line2) in enumerate(zip(f1, f2)): + line1 = line1.strip() + line = line1 + '\t' + line2 + dst.write(line) + + +if __name__ == '__main__': + concat_file('dev.de-en.de', 'dev.de-en.en', 'dev') + concat_file('test.de-en.de', 'test.de-en.en', 'test') + concat_file('train.de-en.de', 'train.de-en.en', 'train') + + src_dict = cPickle.load(open('vocab.de')) + trg_dict = cPickle.load(open('vocab.en')) + + with open('src_dict', 'w') as f: + f.write('\n\nUNK\n') + f.writelines('\n'.join(src_dict.keys())) + + with open('trg_dict', 'w') as f: + f.write('\n\nUNK\n') + f.writelines('\n'.join(trg_dict.keys())) diff --git a/conv_seq2seq/reader.py b/conv_seq2seq/reader.py index 6d4db49f2d34152f4e178034c6e60d287352fe38..ad420af5faade1cd5ee7ef947f7f8920ce6a8bdb 100644 --- a/conv_seq2seq/reader.py +++ b/conv_seq2seq/reader.py @@ -18,7 +18,7 @@ def get_reverse_dict(dictionary): def load_data(data_file, src_dict, trg_dict): - UNK_IDX = src_dict[''] + UNK_IDX = src_dict['UNK'] with open(data_file, 'r') as f: for line in f: line_split = line.strip().split('\t') @@ -34,7 +34,7 @@ def load_data(data_file, src_dict, trg_dict): def data_reader(data_file, src_dict, trg_dict, pos_size, padding_num): def reader(): - UNK_IDX = src_dict[''] + UNK_IDX = src_dict['UNK'] word_padding = trg_dict.__len__() pos_padding = pos_size diff --git a/conv_seq2seq/train.py b/conv_seq2seq/train.py index c6ce0dff12f7b8c3066e791f13c957ef695f1cf5..4bd9a1af675ada5820bb375938a4675e6e71fbe1 100644 --- a/conv_seq2seq/train.py +++ b/conv_seq2seq/train.py @@ -40,7 +40,7 @@ def parse_args(): parser.add_argument( '--emb_size', type=int, - default=512, + default=256, help='Dimension of word embedding. (default: %(default)s)') parser.add_argument( '--pos_size', @@ -52,6 +52,11 @@ def parse_args(): type=float, default=0., help='Dropout rate. (default: %(default)s)') + parser.add_argument( + "--use_bn", + default=False, + type=distutils.util.strtobool, + help="Use batch normalization or not. (default: %(default)s)") parser.add_argument( "--use_gpu", default=False, @@ -116,9 +121,10 @@ def train(train_data_path, trg_dict_path, enc_conv_blocks, dec_conv_blocks, - emb_dim=512, + emb_dim=256, pos_size=200, drop_rate=0., + use_bn=False, batch_size=32, num_passes=15): """ @@ -147,6 +153,8 @@ def train(train_data_path, :type pos_size: int :param drop_rate: Dropout rate. :type drop_rate: float + :param use_bn: Whether to use batch normalization or not. False is the default value. + :type use_bn: bool :param batch_size: The size of a mini-batch. :type batch_size: int :param num_passes: The total number of the passes to train. @@ -158,8 +166,7 @@ def train(train_data_path, src_dict_size = src_dict.__len__() trg_dict_size = trg_dict.__len__() - optimizer = paddle.optimizer.Adam( - learning_rate=1e-3, ) + optimizer = paddle.optimizer.Adam(learning_rate=1e-3, ) cost = conv_seq2seq( src_dict_size=src_dict_size, @@ -169,12 +176,14 @@ def train(train_data_path, enc_conv_blocks=enc_conv_blocks, dec_conv_blocks=dec_conv_blocks, drop_rate=drop_rate, + with_bn=use_bn, is_infer=False) # create parameters and trainer parameters = paddle.parameters.create(cost) - trainer = paddle.trainer.SGD( - cost=cost, parameters=parameters, update_equation=optimizer) + trainer = paddle.trainer.SGD(cost=cost, + parameters=parameters, + update_equation=optimizer) padding_list = [context_len - 1 for (size, context_len) in dec_conv_blocks] padding_num = reduce(lambda x, y: x + y, padding_list) @@ -203,7 +212,6 @@ def train(train_data_path, print "[%s]: Pass: %d, Batch: %d, TrainCost: %f, %s" % ( cur_time, event.pass_id, event.batch_id, event.cost, event.metrics) - else: sys.stdout.flush() if isinstance(event, paddle.event.EndPass): @@ -232,6 +240,8 @@ def main(): enc_conv_blocks = eval(args.enc_blocks) dec_conv_blocks = eval(args.dec_blocks) + sys.setrecursionlimit(10000) + paddle.init(use_gpu=args.use_gpu, trainer_count=args.trainer_count) train( @@ -244,6 +254,7 @@ def main(): emb_dim=args.emb_size, pos_size=args.pos_size, drop_rate=args.drop_rate, + use_bn=args.use_bn, batch_size=args.batch_size, num_passes=args.num_passes) diff --git a/ctr/README.cn.md b/ctr/README.cn.md new file mode 100644 index 0000000000000000000000000000000000000000..a4cb6d17144a9d78a2764e8a49d3abaeb918b7b6 --- /dev/null +++ b/ctr/README.cn.md @@ -0,0 +1,365 @@ +# 点击率预估 + +以下是本例目录包含的文件以及对应说明: + +``` +├── README.md # 本教程markdown 文档 +├── dataset.md # 数据集处理教程 +├── images # 本教程图片目录 +│   ├── lr_vs_dnn.jpg +│   └── wide_deep.png +├── infer.py # 预测脚本 +├── network_conf.py # 模型网络配置 +├── reader.py # data reader +├── train.py # 训练脚本 +└── utils.py # helper functions +└── avazu_data_processer.py # 示例数据预处理脚本 +``` + +## 背景介绍 + +CTR(Click-Through Rate，点击率预估)\[[1](https://en.wikipedia.org/wiki/Click-through_rate)\] +是对用户点击一个特定链接的概率做出预测，是广告投放过程中的一个重要环节。精准的点击率预估对在线广告系统收益最大化具有重要意义。 + +当有多个广告位时，CTR 预估一般会作为排序的基准，比如在搜索引擎的广告系统里，当用户输入一个带商业价值的搜索词（query）时，系统大体上会执行下列步骤来展示广告： + +1. 获取与用户搜索词相关的广告集合 +2. 业务规则和相关性过滤 +3. 根据拍卖机制和 CTR 排序 +4. 展出广告 + +可以看到，CTR 在最终排序中起到了很重要的作用。 + +### 发展阶段 +在业内，CTR 模型经历了如下的发展阶段： + +- Logistic Regression(LR) / GBDT + 特征工程 +- LR + DNN 特征 +- DNN + 特征工程 + +在发展早期时 LR 一统天下，但最近 DNN 模型由于其强大的学习能力和逐渐成熟的性能优化， +逐渐地接过 CTR 预估任务的大旗。 + + +### LR vs DNN + +下图展示了 LR 和一个 \(3x2\) 的 DNN 模型的结构： + +

+
+Figure 1. LR 和 DNN 模型结构对比 +

+ +LR 的蓝色箭头部分可以直接类比到 DNN 中对应的结构，可以看到 LR 和 DNN 有一些共通之处（比如权重累加）， +但前者的模型复杂度在相同输入维度下比后者可能低很多（从某方面讲，模型越复杂，越有潜力学习到更复杂的信息）； +如果 LR 要达到匹敌 DNN 的学习能力，必须增加输入的维度，也就是增加特征的数量， +这也就是为何 LR 和大规模的特征工程必须绑定在一起的原因。 + +LR 对于 DNN 模型的优势是对大规模稀疏特征的容纳能力，包括内存和计算量等方面，工业界都有非常成熟的优化方法； +而 DNN 模型具有自己学习新特征的能力，一定程度上能够提升特征使用的效率， +这使得 DNN 模型在同样规模特征的情况下，更有可能达到更好的学习效果。 + +本文后面的章节会演示如何使用 PaddlePaddle 编写一个结合两者优点的模型。 + + +## 数据和任务抽象 + +我们可以将 `click` 作为学习目标，任务可以有以下几种方案： + +1. 直接学习 click，0,1 作二元分类 +2. Learning to rank, 具体用 pairwise rank（标签 1>0）或者 listwise rank +3. 统计每个广告的点击率，将同一个 query 下的广告两两组合，点击率高的>点击率低的，做 rank 或者分类 + +我们直接使用第一种方法做分类任务。 + +我们使用 Kaggle 上 `Click-through rate prediction` 任务的数据集\[[2](https://www.kaggle.com/c/avazu-ctr-prediction/data)\] 来演示本例中的模型。 + +具体的特征处理方法参看 [data process](./dataset.md)。 + +本教程中演示模型的输入格式如下： + +``` +# \t \t click +1 23 190 \t 230:0.12 3421:0.9 23451:0.12 \t 0 +23 231 \t 1230:0.12 13421:0.9 \t 1 +``` + +详细的格式描述如下： + +- `dnn input ids` 采用 one-hot 表示，只需要填写值为1的ID（注意这里不是变长输入） +- `lr input sparse values` 使用了 `ID:VALUE` 的表示，值部分最好规约到值域 `[-1, 1]`。 + +此外，模型训练时需要传入一个文件描述 dnn 和 lr两个子模型的输入维度，文件的格式如下： + +``` +dnn_input_dim: +lr_input_dim: +``` + +其中， `` 表示一个整型数值。 + +本目录下的 `avazu_data_processor.py` 可以对下载的演示数据集\[[2](#参考文档)\] 进行处理，具体使用方法参考如下说明： + +``` +usage: avazu_data_processer.py [-h] --data_path DATA_PATH --output_dir + OUTPUT_DIR + [--num_lines_to_detect NUM_LINES_TO_DETECT] + [--test_set_size TEST_SET_SIZE] + [--train_size TRAIN_SIZE] + +PaddlePaddle CTR example + +optional arguments: + -h, --help show this help message and exit + --data_path DATA_PATH + path of the Avazu dataset + --output_dir OUTPUT_DIR + directory to output + --num_lines_to_detect NUM_LINES_TO_DETECT + number of records to detect dataset's meta info + --test_set_size TEST_SET_SIZE + size of the validation dataset(default: 10000) + --train_size TRAIN_SIZE + size of the trainset (default: 100000) +``` + +- `data_path` 是待处理的数据路径 +- `output_dir` 生成数据的输出路径 +- `num_lines_to_detect` 预先扫描数据生成ID的个数，这里是扫描的文件行数 +- `test_set_size` 生成测试集的行数 +- `train_size` 生成训练姐的行数 + +## Wide & Deep Learning Model + +谷歌在 16 年提出了 Wide & Deep Learning 的模型框架，用于融合适合学习抽象特征的 DNN 和 适用于大规模稀疏特征的 LR 两种模型的优点。 + + +### 模型简介 + +Wide & Deep Learning Model\[[3](#参考文献)\] 可以作为一种相对成熟的模型框架使用， +在 CTR 预估的任务中工业界也有一定的应用，因此本文将演示使用此模型来完成 CTR 预估的任务。 + +模型结构如下： + +

+
+Figure 2. Wide & Deep Model +

+ +模型上边的 Wide 部分，可以容纳大规模系数特征，并且对一些特定的信息（比如 ID）有一定的记忆能力； +而模型下边的 Deep 部分，能够学习特征间的隐含关系，在相同数量的特征下有更好的学习和推导能力。 + + +### 编写模型输入 + +模型只接受 3 个输入，分别是 + +- `dnn_input` ，也就是 Deep 部分的输入 +- `lr_input` ，也就是 Wide 部分的输入 +- `click` ， 点击与否，作为二分类模型学习的标签 + +```python +dnn_merged_input = layer.data( + name='dnn_input', + type=paddle.data_type.sparse_binary_vector(data_meta_info['dnn_input'])) + +lr_merged_input = layer.data( + name='lr_input', + type=paddle.data_type.sparse_binary_vector(data_meta_info['lr_input'])) + +click = paddle.layer.data(name='click', type=dtype.dense_vector(1)) +``` + +### 编写 Wide 部分 + +Wide 部分直接使用了 LR 模型，但激活函数改成了 `RELU` 来加速 + +```python +def build_lr_submodel(): + fc = layer.fc( + input=lr_merged_input, size=1, name='lr', act=paddle.activation.Relu()) + return fc +``` + +### 编写 Deep 部分 + +Deep 部分使用了标准的多层前向传导的 DNN 模型 + +```python +def build_dnn_submodel(dnn_layer_dims): + dnn_embedding = layer.fc(input=dnn_merged_input, size=dnn_layer_dims[0]) + _input_layer = dnn_embedding + for i, dim in enumerate(dnn_layer_dims[1:]): + fc = layer.fc( + input=_input_layer, + size=dim, + act=paddle.activation.Relu(), + name='dnn-fc-%d' % i) + _input_layer = fc + return _input_layer +``` + +### 两者融合 + +两个 submodel 的最上层输出加权求和得到整个模型的输出，输出部分使用 `sigmoid` 作为激活函数，得到区间 (0,1) 的预测值， +来逼近训练数据中二元类别的分布，并最终作为 CTR 预估的值使用。 + +```python +# conbine DNN and LR submodels +def combine_submodels(dnn, lr): + merge_layer = layer.concat(input=[dnn, lr]) + fc = layer.fc( + input=merge_layer, + size=1, + name='output', + # use sigmoid function to approximate ctr, wihch is a float value between 0 and 1. + act=paddle.activation.Sigmoid()) + return fc +``` + +### 训练任务的定义 +```python +dnn = build_dnn_submodel(dnn_layer_dims) +lr = build_lr_submodel() +output = combine_submodels(dnn, lr) + +# ============================================================================== +# cost and train period +# ============================================================================== +classification_cost = paddle.layer.multi_binary_label_cross_entropy_cost( + input=output, label=click) + + +paddle.init(use_gpu=False, trainer_count=11) + +params = paddle.parameters.create(classification_cost) + +optimizer = paddle.optimizer.Momentum(momentum=0) + +trainer = paddle.trainer.SGD( + cost=classification_cost, parameters=params, update_equation=optimizer) + +dataset = AvazuDataset(train_data_path, n_records_as_test=test_set_size) + +def event_handler(event): + if isinstance(event, paddle.event.EndIteration): + if event.batch_id % 100 == 0: + logging.warning("Pass %d, Samples %d, Cost %f" % ( + event.pass_id, event.batch_id * batch_size, event.cost)) + + if event.batch_id % 1000 == 0: + result = trainer.test( + reader=paddle.batch(dataset.test, batch_size=1000), + feeding=field_index) + logging.warning("Test %d-%d, Cost %f" % (event.pass_id, event.batch_id, + result.cost)) + + +trainer.train( + reader=paddle.batch( + paddle.reader.shuffle(dataset.train, buf_size=500), + batch_size=batch_size), + feeding=field_index, + event_handler=event_handler, + num_passes=100) +``` +## 运行训练和测试 +训练模型需要如下步骤： + +1. 准备训练数据 + 1. 从 [Kaggle CTR](https://www.kaggle.com/c/avazu-ctr-prediction/data) 下载 train.gz + 2. 解压 train.gz 得到 train.txt + 3. `mkdir -p output; python avazu_data_processer.py --data_path train.txt --output_dir output --num_lines_to_detect 1000 --test_set_size 100` 生成演示数据 +2. 执行 `python train.py --train_data_path ./output/train.txt --test_data_path ./output/test.txt --data_meta_file ./output/data.meta.txt --model_type=0` 开始训练 + +上面第2个步骤可以为 `train.py` 填充命令行参数来定制模型的训练过程，具体的命令行参数及用法如下 + +``` +usage: train.py [-h] --train_data_path TRAIN_DATA_PATH + [--test_data_path TEST_DATA_PATH] [--batch_size BATCH_SIZE] + [--num_passes NUM_PASSES] + [--model_output_prefix MODEL_OUTPUT_PREFIX] --data_meta_file + DATA_META_FILE --model_type MODEL_TYPE + +PaddlePaddle CTR example + +optional arguments: + -h, --help show this help message and exit + --train_data_path TRAIN_DATA_PATH + path of training dataset + --test_data_path TEST_DATA_PATH + path of testing dataset + --batch_size BATCH_SIZE + size of mini-batch (default:10000) + --num_passes NUM_PASSES + number of passes to train + --model_output_prefix MODEL_OUTPUT_PREFIX + prefix of path for model to store (default: + ./ctr_models) + --data_meta_file DATA_META_FILE + path of data meta info file + --model_type MODEL_TYPE + model type, classification: 0, regression 1 (default + classification) +``` + +- `train_data_path` ： 训练集的路径 +- `test_data_path` : 测试集的路径 +- `num_passes`: 模型训练多少轮 +- `data_meta_file`: 参考[数据和任务抽象](### 数据和任务抽象)的描述。 +- `model_type`: 模型分类或回归 + + +## 用训好的模型做预测 +训好的模型可以用来预测新的数据， 预测数据的格式为 + +``` +# \t +1 23 190 \t 230:0.12 3421:0.9 23451:0.12 +23 231 \t 1230:0.12 13421:0.9 +``` + +这里与训练数据的格式唯一不同的地方，就是没有标签，也就是训练数据中第3列 `click` 对应的数值。 + +`infer.py` 的使用方法如下 + +``` +usage: infer.py [-h] --model_gz_path MODEL_GZ_PATH --data_path DATA_PATH + --prediction_output_path PREDICTION_OUTPUT_PATH + [--data_meta_path DATA_META_PATH] --model_type MODEL_TYPE + +PaddlePaddle CTR example + +optional arguments: + -h, --help show this help message and exit + --model_gz_path MODEL_GZ_PATH + path of model parameters gz file + --data_path DATA_PATH + path of the dataset to infer + --prediction_output_path PREDICTION_OUTPUT_PATH + path to output the prediction + --data_meta_path DATA_META_PATH + path of trainset's meta info, default is ./data.meta + --model_type MODEL_TYPE + model type, classification: 0, regression 1 (default + classification) +``` + +- `model_gz_path_model`：用 `gz` 压缩过的模型路径 +- `data_path` ： 需要预测的数据路径 +- `prediction_output_paht`：预测输出的路径 +- `data_meta_file` ：参考[数据和任务抽象](### 数据和任务抽象)的描述。 +- `model_type` ：分类或回归 + +示例数据可以用如下命令预测 + +``` +python infer.py --model_gz_path --data_path output/infer.txt --prediction_output_path predictions.txt --data_meta_path data.meta.txt +``` + +最终的预测结果位于 `predictions.txt`。 + +## 参考文献 +1. +2. +3. Cheng H T, Koc L, Harmsen J, et al. [Wide & deep learning for recommender systems](https://arxiv.org/pdf/1606.07792.pdf)[C]//Proceedings of the 1st Workshop on Deep Learning for Recommender Systems. ACM, 2016: 7-10. diff --git a/ctr/README.md b/ctr/README.md index 391c842edb4e4d16badd23a4f3c2545dbb16df89..6f11ac60734b4a549a9c84d7fbba8ed283a97284 100644 --- a/ctr/README.md +++ b/ctr/README.md @@ -120,7 +120,7 @@ The model structure is as follows: Figure 2. Wide & Deep Model

-The wide part of the left side of the model can accommodate large-scale coefficient features and has some memory for some specific information (such as ID); and the Deep part of the right side of the model can learn the implicit relationship between features. +The wide part of the top side of the model can accommodate large-scale coefficient features and has some memory for some specific information (such as ID); and the Deep part of the bottom side of the model can learn the implicit relationship between features. ### Model Input diff --git a/ctr/avazu_data_processer.py b/ctr/avazu_data_processer.py index 18aa85330ae765a236416ab1826092c1fe6a133a..dd3c1441f8f8b26473d15889198abb3593edfa51 100644 --- a/ctr/avazu_data_processer.py +++ b/ctr/avazu_data_processer.py @@ -79,8 +79,9 @@ all the files are for demo. feature_dims = {} -categorial_features = ('C1 banner_pos site_category app_category ' + - 'device_type device_conn_type').split() +categorial_features = ( + 'C1 banner_pos site_category app_category ' + 'device_type device_conn_type' +).split() id_features = 'id site_id app_id device_id _device_id_cross_site_id'.split() @@ -335,8 +336,8 @@ class AvazuDataset(object): else: fea0 = self.fields[key].cross_fea0 fea1 = self.fields[key].cross_fea1 - record.append( - self.fields[key].gen_cross_fea(row[fea0], row[fea1])) + record.append(self.fields[key].gen_cross_fea(row[fea0], row[ + fea1])) sparse_input = concat_sparse_vectors(record, self.id_dims) @@ -396,8 +397,9 @@ with open(output_infer_path, 'w') as f: dnn_input, lr_input = record dnn_input = ids2dense(dnn_input, feature_dims['dnn_input']) lr_input = ids2sparse(lr_input) - line = "%s\t%s\n" % (' '.join(map(str, dnn_input)), - ' '.join(map(str, lr_input)), ) + line = "%s\t%s\n" % ( + ' '.join(map(str, dnn_input)), + ' '.join(map(str, lr_input)), ) f.write(line) if id > args.test_set_size: break diff --git a/ctr/images/wide_deep.png b/ctr/images/wide_deep.png index 03c4afcfc6cea0b5abf4c4554ecf9810843e75e2..616f88cb22607c1c6bcbe4312644f632ef284e8e 100644 Binary files a/ctr/images/wide_deep.png and b/ctr/images/wide_deep.png differ diff --git a/ctr/network_conf.py b/ctr/network_conf.py index f6f4e4a59e17114b279149048c3bf144f2987faa..bcff49ee05e1d8cc80e2fdd28a771bf9bf9502e3 100644 --- a/ctr/network_conf.py +++ b/ctr/network_conf.py @@ -60,15 +60,14 @@ class CTRmodel(object): ''' build DNN submodel. ''' - dnn_embedding = layer.fc( - input=self.dnn_merged_input, size=dnn_layer_dims[0]) + dnn_embedding = layer.fc(input=self.dnn_merged_input, + size=dnn_layer_dims[0]) _input_layer = dnn_embedding for i, dim in enumerate(dnn_layer_dims[1:]): - fc = layer.fc( - input=_input_layer, - size=dim, - act=paddle.activation.Relu(), - name='dnn-fc-%d' % i) + fc = layer.fc(input=_input_layer, + size=dim, + act=paddle.activation.Relu(), + name='dnn-fc-%d' % i) _input_layer = fc return _input_layer @@ -76,8 +75,9 @@ class CTRmodel(object): ''' config LR submodel ''' - fc = layer.fc( - input=self.lr_merged_input, size=1, act=paddle.activation.Relu()) + fc = layer.fc(input=self.lr_merged_input, + size=1, + act=paddle.activation.Relu()) return fc def _build_classification_model(self, dnn, lr): @@ -95,8 +95,9 @@ class CTRmodel(object): def _build_regression_model(self, dnn, lr): merge_layer = layer.concat(input=[dnn, lr]) - self.output = layer.fc( - input=merge_layer, size=1, act=paddle.activation.Sigmoid()) + self.output = layer.fc(input=merge_layer, + size=1, + act=paddle.activation.Sigmoid()) if not self.is_infer: self.train_cost = paddle.layer.square_error_cost( input=self.output, label=self.click) diff --git a/ctr/train.py b/ctr/train.py index 421425e2a701e3c413e23ab81930e3c7e68d40ef..de7add61d65aba363cc17bed49d32c9054600108 100644 --- a/ctr/train.py +++ b/ctr/train.py @@ -68,8 +68,9 @@ def train(): params = paddle.parameters.create(model.train_cost) optimizer = paddle.optimizer.AdaGrad() - trainer = paddle.trainer.SGD( - cost=model.train_cost, parameters=params, update_equation=optimizer) + trainer = paddle.trainer.SGD(cost=model.train_cost, + parameters=params, + update_equation=optimizer) dataset = reader.Dataset() diff --git a/ctr/utils.py b/ctr/utils.py index d8cf569cc9b28fa04ee389838cef9edd6862e52b..437554c3c291d5a74cc0b3844c8684c73b189a19 100644 --- a/ctr/utils.py +++ b/ctr/utils.py @@ -64,5 +64,7 @@ def load_dnn_input_record(sent): def load_lr_input_record(sent): res = [] for _ in [x.split(':') for x in sent.split()]: - res.append((int(_[0]), float(_[1]), )) + res.append(( + int(_[0]), + float(_[1]), )) return res diff --git a/deep_fm/README.md b/deep_fm/README.md new file mode 100644 index 0000000000000000000000000000000000000000..aa63170c921e7a22801790834a0db86df5e70e7a --- /dev/null +++ b/deep_fm/README.md @@ -0,0 +1,91 @@ +# Deep Factorization Machine for Click-Through Rate prediction + +## Introduction +This model implements the DeepFM proposed in the following paper: + +```text +@inproceedings{guo2017deepfm, + title={DeepFM: A Factorization-Machine based Neural Network for CTR Prediction}, + author={Huifeng Guo, Ruiming Tang, Yunming Ye, Zhenguo Li and Xiuqiang He}, + booktitle={the Twenty-Sixth International Joint Conference on Artificial Intelligence (IJCAI)}, + pages={1725--1731}, + year={2017} +} +``` + +The DeepFm combines factorization machine and deep neural networks to model +both low order and high order feature interactions. For details of the +factorization machines, please refer to the paper [factorization +machines](https://www.csie.ntu.edu.tw/~b97053/paper/Rendle2010FM.pdf) + +## Dataset +This example uses Criteo dataset which was used for the [Display Advertising +Challenge](https://www.kaggle.com/c/criteo-display-ad-challenge/) +hosted by Kaggle. + +Each row is the features for an ad display and the first column is a label +indicating whether this ad has been clicked or not. There are 39 features in +total. 13 features take integer values and the other 26 features are +categorical features. For the test dataset, the labels are omitted. + +Download dataset: +```bash +cd data && ./download.sh && cd .. +``` + +## Model +The DeepFM model is composed of the factorization machine layer (FM) and deep +neural networks (DNN). All the input features are feeded to both FM and DNN. +The output from FM and DNN are combined to form the final output. The embedding +layer for sparse features in the DNN shares the parameters with the latent +vectors (factors) of the FM layer. + +The factorization machine layer in PaddlePaddle computes the second order +interactions. The following code example combines the factorization machine +layer and fully connected layer to form the full version of factorization +machine: + +```python +def fm_layer(input, factor_size): + first_order = paddle.layer.fc(input=input, size=1, act=paddle.activation.Linear()) + second_order = paddle.layer.factorization_machine(input=input, factor_size=factor_size) + fm = paddle.layer.addto(input=[first_order, second_order], + act=paddle.activation.Linear(), + bias_attr=False) + return fm +``` + +## Data preparation +To preprocess the raw dataset, the integer features are clipped then min-max +normalized to [0, 1] and the categorical features are one-hot encoded. The raw +training dataset are splited such that 90% are used for training and the other +10% are used for validation during training. + +```bash +python preprocess.py --datadir ./data/raw --outdir ./data +``` + +## Train +The command line options for training can be listed by `python train.py -h`. + +To train the model: +```bash +python train.py \ + --train_data_path data/train.txt \ + --test_data_path data/valid.txt \ + 2>&1 | tee train.log +``` + +After training pass 9 batch 40000, the testing AUC is `0.807178` and the testing +cost is `0.445196`. + +## Infer +The command line options for infering can be listed by `python infer.py -h`. + +To make inference for the test dataset: +```bash +python infer.py \ + --model_gz_path models/model-pass-9-batch-10000.tar.gz \ + --data_path data/test.txt \ + --prediction_output_path ./predict.txt +``` diff --git a/deep_fm/data/download.sh b/deep_fm/data/download.sh new file mode 100755 index 0000000000000000000000000000000000000000..466a22f2c6cc885cea0a1468f3043cb59c611b59 --- /dev/null +++ b/deep_fm/data/download.sh @@ -0,0 +1,8 @@ +#!/bin/bash + +wget --no-check-certificate https://s3-eu-west-1.amazonaws.com/criteo-labs/dac.tar.gz +tar zxf dac.tar.gz +rm -f dac.tar.gz + +mkdir raw +mv ./*.txt raw/ diff --git a/deep_fm/infer.py b/deep_fm/infer.py new file mode 100755 index 0000000000000000000000000000000000000000..40a5929780090d403b8b905f8e949f1f8a020eb3 --- /dev/null +++ b/deep_fm/infer.py @@ -0,0 +1,63 @@ +import os +import gzip +import argparse +import itertools + +import paddle.v2 as paddle + +from network_conf import DeepFM +import reader + + +def parse_args(): + parser = argparse.ArgumentParser(description="PaddlePaddle DeepFM example") + parser.add_argument( + '--model_gz_path', + type=str, + required=True, + help="The path of model parameters gz file") + parser.add_argument( + '--data_path', + type=str, + required=True, + help="The path of the dataset to infer") + parser.add_argument( + '--prediction_output_path', + type=str, + required=True, + help="The path to output the prediction") + parser.add_argument( + '--factor_size', + type=int, + default=10, + help="The factor size for the factorization machine (default:10)") + + return parser.parse_args() + + +def infer(): + args = parse_args() + + paddle.init(use_gpu=False, trainer_count=1) + + model = DeepFM(args.factor_size, infer=True) + + parameters = paddle.parameters.Parameters.from_tar( + gzip.open(args.model_gz_path, 'r')) + + inferer = paddle.inference.Inference( + output_layer=model, parameters=parameters) + + dataset = reader.Dataset() + + infer_reader = paddle.batch(dataset.infer(args.data_path), batch_size=1000) + + with open(args.prediction_output_path, 'w') as out: + for id, batch in enumerate(infer_reader()): + res = inferer.infer(input=batch) + predictions = [x for x in itertools.chain.from_iterable(res)] + out.write('\n'.join(map(str, predictions)) + '\n') + + +if __name__ == '__main__': + infer() diff --git a/deep_fm/network_conf.py b/deep_fm/network_conf.py new file mode 100644 index 0000000000000000000000000000000000000000..545fe07b8197e3379eb5a6f34c3134b813a4684e --- /dev/null +++ b/deep_fm/network_conf.py @@ -0,0 +1,75 @@ +import paddle.v2 as paddle + +dense_feature_dim = 13 +sparse_feature_dim = 117568 + + +def fm_layer(input, factor_size, fm_param_attr): + first_order = paddle.layer.fc(input=input, + size=1, + act=paddle.activation.Linear()) + second_order = paddle.layer.factorization_machine( + input=input, + factor_size=factor_size, + act=paddle.activation.Linear(), + param_attr=fm_param_attr) + out = paddle.layer.addto( + input=[first_order, second_order], + act=paddle.activation.Linear(), + bias_attr=False) + return out + + +def DeepFM(factor_size, infer=False): + dense_input = paddle.layer.data( + name="dense_input", + type=paddle.data_type.dense_vector(dense_feature_dim)) + sparse_input = paddle.layer.data( + name="sparse_input", + type=paddle.data_type.sparse_binary_vector(sparse_feature_dim)) + sparse_input_ids = [ + paddle.layer.data( + name="C" + str(i), + type=paddle.data_type.integer_value(sparse_feature_dim)) + for i in range(1, 27) + ] + + dense_fm = fm_layer( + dense_input, + factor_size, + fm_param_attr=paddle.attr.Param(name="DenseFeatFactors")) + sparse_fm = fm_layer( + sparse_input, + factor_size, + fm_param_attr=paddle.attr.Param(name="SparseFeatFactors")) + + def embedding_layer(input): + return paddle.layer.embedding( + input=input, + size=factor_size, + param_attr=paddle.attr.Param(name="SparseFeatFactors")) + + sparse_embed_seq = map(embedding_layer, sparse_input_ids) + sparse_embed = paddle.layer.concat(sparse_embed_seq) + + fc1 = paddle.layer.fc(input=[sparse_embed, dense_input], + size=400, + act=paddle.activation.Relu()) + fc2 = paddle.layer.fc(input=fc1, size=400, act=paddle.activation.Relu()) + fc3 = paddle.layer.fc(input=fc2, size=400, act=paddle.activation.Relu()) + + predict = paddle.layer.fc(input=[dense_fm, sparse_fm, fc3], + size=1, + act=paddle.activation.Sigmoid()) + + if not infer: + label = paddle.layer.data( + name="label", type=paddle.data_type.dense_vector(1)) + cost = paddle.layer.multi_binary_label_cross_entropy_cost( + input=predict, label=label) + paddle.evaluator.classification_error( + name="classification_error", input=predict, label=label) + paddle.evaluator.auc(name="auc", input=predict, label=label) + return cost + else: + return predict diff --git a/deep_fm/preprocess.py b/deep_fm/preprocess.py new file mode 100755 index 0000000000000000000000000000000000000000..36ffea16637c19dee9352d17ed51a67edf582167 --- /dev/null +++ b/deep_fm/preprocess.py @@ -0,0 +1,164 @@ +""" +Preprocess Criteo dataset. This dataset was used for the Display Advertising +Challenge (https://www.kaggle.com/c/criteo-display-ad-challenge). +""" +import os +import sys +import click +import random +import collections + +# There are 13 integer features and 26 categorical features +continous_features = range(1, 14) +categorial_features = range(14, 40) + +# Clip integer features. The clip point for each integer feature +# is derived from the 95% quantile of the total values in each feature +continous_clip = [20, 600, 100, 50, 64000, 500, 100, 50, 500, 10, 10, 10, 50] + + +class CategoryDictGenerator: + """ + Generate dictionary for each of the categorical features + """ + + def __init__(self, num_feature): + self.dicts = [] + self.num_feature = num_feature + for i in range(0, num_feature): + self.dicts.append(collections.defaultdict(int)) + + def build(self, datafile, categorial_features, cutoff=0): + with open(datafile, 'r') as f: + for line in f: + features = line.rstrip('\n').split('\t') + for i in range(0, self.num_feature): + if features[categorial_features[i]] != '': + self.dicts[i][features[categorial_features[i]]] += 1 + for i in range(0, self.num_feature): + self.dicts[i] = filter(lambda x: x[1] >= cutoff, + self.dicts[i].items()) + self.dicts[i] = sorted(self.dicts[i], key=lambda x: (-x[1], x[0])) + vocabs, _ = list(zip(*self.dicts[i])) + self.dicts[i] = dict(zip(vocabs, range(1, len(vocabs) + 1))) + self.dicts[i][''] = 0 + + def gen(self, idx, key): + if key not in self.dicts[idx]: + res = self.dicts[idx][''] + else: + res = self.dicts[idx][key] + return res + + def dicts_sizes(self): + return map(len, self.dicts) + + +class ContinuousFeatureGenerator: + """ + Normalize the integer features to [0, 1] by min-max normalization + """ + + def __init__(self, num_feature): + self.num_feature = num_feature + self.min = [sys.maxint] * num_feature + self.max = [-sys.maxint] * num_feature + + def build(self, datafile, continous_features): + with open(datafile, 'r') as f: + for line in f: + features = line.rstrip('\n').split('\t') + for i in range(0, self.num_feature): + val = features[continous_features[i]] + if val != '': + val = int(val) + if val > continous_clip[i]: + val = continous_clip[i] + self.min[i] = min(self.min[i], val) + self.max[i] = max(self.max[i], val) + + def gen(self, idx, val): + if val == '': + return 0.0 + val = float(val) + return (val - self.min[idx]) / (self.max[idx] - self.min[idx]) + + +@click.command("preprocess") +@click.option("--datadir", type=str, help="Path to raw criteo dataset") +@click.option("--outdir", type=str, help="Path to save the processed data") +def preprocess(datadir, outdir): + """ + All the 13 integer features are normalzied to continous values and these + continous features are combined into one vecotr with dimension 13. + + Each of the 26 categorical features are one-hot encoded and all the one-hot + vectors are combined into one sparse binary vector. + """ + dists = ContinuousFeatureGenerator(len(continous_features)) + dists.build(os.path.join(datadir, 'train.txt'), continous_features) + + dicts = CategoryDictGenerator(len(categorial_features)) + dicts.build( + os.path.join(datadir, 'train.txt'), categorial_features, cutoff=200) + + dict_sizes = dicts.dicts_sizes() + categorial_feature_offset = [0] + for i in range(1, len(categorial_features)): + offset = categorial_feature_offset[i - 1] + dict_sizes[i - 1] + categorial_feature_offset.append(offset) + + random.seed(0) + + # 90% of the data are used for training, and 10% of the data are used + # for validation. + with open(os.path.join(outdir, 'train.txt'), 'w') as out_train: + with open(os.path.join(outdir, 'valid.txt'), 'w') as out_valid: + with open(os.path.join(datadir, 'train.txt'), 'r') as f: + for line in f: + features = line.rstrip('\n').split('\t') + + continous_vals = [] + for i in range(0, len(continous_features)): + val = dists.gen(i, features[continous_features[i]]) + continous_vals.append("{0:.6f}".format(val).rstrip('0') + .rstrip('.')) + categorial_vals = [] + for i in range(0, len(categorial_features)): + val = dicts.gen(i, features[categorial_features[ + i]]) + categorial_feature_offset[i] + categorial_vals.append(str(val)) + + continous_vals = ','.join(continous_vals) + categorial_vals = ','.join(categorial_vals) + label = features[0] + if random.randint(0, 9999) % 10 != 0: + out_train.write('\t'.join( + [continous_vals, categorial_vals, label]) + '\n') + else: + out_valid.write('\t'.join( + [continous_vals, categorial_vals, label]) + '\n') + + with open(os.path.join(outdir, 'test.txt'), 'w') as out: + with open(os.path.join(datadir, 'test.txt'), 'r') as f: + for line in f: + features = line.rstrip('\n').split('\t') + + continous_vals = [] + for i in range(0, len(continous_features)): + val = dists.gen(i, features[continous_features[i] - 1]) + continous_vals.append("{0:.6f}".format(val).rstrip('0') + .rstrip('.')) + categorial_vals = [] + for i in range(0, len(categorial_features)): + val = dicts.gen(i, features[categorial_features[ + i] - 1]) + categorial_feature_offset[i] + categorial_vals.append(str(val)) + + continous_vals = ','.join(continous_vals) + categorial_vals = ','.join(categorial_vals) + out.write('\t'.join([continous_vals, categorial_vals]) + '\n') + + +if __name__ == "__main__": + preprocess() diff --git a/deep_fm/reader.py b/deep_fm/reader.py new file mode 100644 index 0000000000000000000000000000000000000000..1098ce423c9071864671be91dea81972e47fbc98 --- /dev/null +++ b/deep_fm/reader.py @@ -0,0 +1,58 @@ +class Dataset: + def _reader_creator(self, path, is_infer): + def reader(): + with open(path, 'r') as f: + for line in f: + features = line.rstrip('\n').split('\t') + dense_feature = map(float, features[0].split(',')) + sparse_feature = map(int, features[1].split(',')) + if not is_infer: + label = [float(features[2])] + yield [dense_feature, sparse_feature + ] + sparse_feature + [label] + else: + yield [dense_feature, sparse_feature] + sparse_feature + + return reader + + def train(self, path): + return self._reader_creator(path, False) + + def test(self, path): + return self._reader_creator(path, False) + + def infer(self, path): + return self._reader_creator(path, True) + + +feeding = { + 'dense_input': 0, + 'sparse_input': 1, + 'C1': 2, + 'C2': 3, + 'C3': 4, + 'C4': 5, + 'C5': 6, + 'C6': 7, + 'C7': 8, + 'C8': 9, + 'C9': 10, + 'C10': 11, + 'C11': 12, + 'C12': 13, + 'C13': 14, + 'C14': 15, + 'C15': 16, + 'C16': 17, + 'C17': 18, + 'C18': 19, + 'C19': 20, + 'C20': 21, + 'C21': 22, + 'C22': 23, + 'C23': 24, + 'C24': 25, + 'C25': 26, + 'C26': 27, + 'label': 28 +} diff --git a/deep_fm/train.py b/deep_fm/train.py new file mode 100755 index 0000000000000000000000000000000000000000..92d48696d8845ac13b714b66f7810acdd35fe164 --- /dev/null +++ b/deep_fm/train.py @@ -0,0 +1,108 @@ +import os +import gzip +import logging +import argparse + +import paddle.v2 as paddle + +from network_conf import DeepFM +import reader + +logging.basicConfig() +logger = logging.getLogger("paddle") +logger.setLevel(logging.INFO) + + +def parse_args(): + parser = argparse.ArgumentParser(description="PaddlePaddle DeepFM example") + parser.add_argument( + '--train_data_path', + type=str, + required=True, + help="The path of training dataset") + parser.add_argument( + '--test_data_path', + type=str, + required=True, + help="The path of testing dataset") + parser.add_argument( + '--batch_size', + type=int, + default=1000, + help="The size of mini-batch (default:1000)") + parser.add_argument( + '--num_passes', + type=int, + default=10, + help="The number of passes to train (default: 10)") + parser.add_argument( + '--factor_size', + type=int, + default=10, + help="The factor size for the factorization machine (default:10)") + parser.add_argument( + '--model_output_dir', + type=str, + default='models', + help='The path for model to store (default: models)') + + return parser.parse_args() + + +def train(): + args = parse_args() + + if not os.path.isdir(args.model_output_dir): + os.mkdir(args.model_output_dir) + + paddle.init(use_gpu=False, trainer_count=1) + + optimizer = paddle.optimizer.Adam(learning_rate=1e-4) + + model = DeepFM(args.factor_size) + + params = paddle.parameters.create(model) + + trainer = paddle.trainer.SGD(cost=model, + parameters=params, + update_equation=optimizer) + + dataset = reader.Dataset() + + def __event_handler__(event): + if isinstance(event, paddle.event.EndIteration): + num_samples = event.batch_id * args.batch_size + if event.batch_id % 100 == 0: + logger.warning("Pass %d, Batch %d, Samples %d, Cost %f, %s" % + (event.pass_id, event.batch_id, num_samples, + event.cost, event.metrics)) + + if event.batch_id % 10000 == 0: + if args.test_data_path: + result = trainer.test( + reader=paddle.batch( + dataset.test(args.test_data_path), + batch_size=args.batch_size), + feeding=reader.feeding) + logger.warning("Test %d-%d, Cost %f, %s" % + (event.pass_id, event.batch_id, result.cost, + result.metrics)) + + path = "{}/model-pass-{}-batch-{}.tar.gz".format( + args.model_output_dir, event.pass_id, event.batch_id) + with gzip.open(path, 'w') as f: + trainer.save_parameter_to_tar(f) + + trainer.train( + reader=paddle.batch( + paddle.reader.shuffle( + dataset.train(args.train_data_path), + buf_size=args.batch_size * 10000), + batch_size=args.batch_size), + feeding=reader.feeding, + event_handler=__event_handler__, + num_passes=args.num_passes) + + +if __name__ == '__main__': + train() diff --git a/deep_speech_2/README.md b/deep_speech_2/README.md deleted file mode 100644 index 427331fcc7bde2a277b6d2cd8314c0d8dd9997df..0000000000000000000000000000000000000000 --- a/deep_speech_2/README.md +++ /dev/null @@ -1,535 +0,0 @@ -Deprecated: please check out the new repository [DeepSpeech](https://github.com/PaddlePaddle/DeepSpeech). - -# DeepSpeech2 on PaddlePaddle - -*DeepSpeech2 on PaddlePaddle* is an open-source implementation of end-to-end Automatic Speech Recognition (ASR) engine, based on [Baidu's Deep Speech 2 paper](http://proceedings.mlr.press/v48/amodei16.pdf), with [PaddlePaddle](https://github.com/PaddlePaddle/Paddle) platform. Our vision is to empower both industrial application and academic research on speech recognition, via an easy-to-use, efficient and scalable implementation, including training, inference & testing module, distributed [PaddleCloud](https://github.com/PaddlePaddle/cloud) training, and demo deployment. Besides, several pre-trained models for both English and Mandarin are also released. - -## Table of Contents -- [Installation](#installation) -- [Getting Started](#getting-started) -- [Data Preparation](#data-preparation) -- [Training a Model](#training-a-model) -- [Data Augmentation Pipeline](#data-augmentation-pipeline) -- [Inference and Evaluation](#inference-and-evaluation) -- [Running in Docker Container](#running-in-docker-container) -- [Distributed Cloud Training](#distributed-cloud-training) -- [Hyper-parameters Tuning](#hyper-parameters-tuning) -- [Training for Mandarin Language](#training-for-mandarin-language) -- [Trying Live Demo with Your Own Voice](#trying-live-demo-with-your-own-voice) -- [Released Models](#released-models) -- [Experiments and Benchmarks](#experiments-and-benchmarks) -- [Questions and Help](#questions-and-help) - - - -## Installation - -To avoid the trouble of environment setup, [running in docker container](#running-in-docker-container) is highly recommended. Otherwise follow the guidelines below to install the dependencies manually. - -### Prerequisites -- Python 2.7 only supported -- PaddlePaddle the latest version (please refer to the [Installation Guide](https://github.com/PaddlePaddle/Paddle#installation)) - -### Setup - -```bash -git clone https://github.com/PaddlePaddle/models.git -cd models/deep_speech_2 -sh setup.sh -``` - -## Getting Started - -Several shell scripts provided in `./examples` will help us to quickly give it a try, for most major modules, including data preparation, model training, case inference and model evaluation, with a few public dataset (e.g. [LibriSpeech](http://www.openslr.org/12/), [Aishell](http://www.openslr.org/33)). Reading these examples will also help you to understand how to make it work with your own data. - -Some of the scripts in `./examples` are configured with 8 GPUs. If you don't have 8 GPUs available, please modify `CUDA_VISIBLE_DEVICES` and `--trainer_count`. If you don't have any GPU available, please set `--use_gpu` to False to use CPUs instead. Besides, if out-of-memory problem occurs, just reduce `--batch_size` to fit. - -Let's take a tiny sampled subset of [LibriSpeech dataset](http://www.openslr.org/12/) for instance. - -- Go to directory - - ```bash - cd examples/tiny - ``` - - Notice that this is only a toy example with a tiny sampled subset of LibriSpeech. If you would like to try with the complete dataset (would take several days for training), please go to `examples/librispeech` instead. -- Prepare the data - - ```bash - sh run_data.sh - ``` - - `run_data.sh` will download dataset, generate manifests, collect normalizer's statistics and build vocabulary. Once the data preparation is done, you will find the data (only part of LibriSpeech) downloaded in `~/.cache/paddle/dataset/speech/libri` and the corresponding manifest files generated in `./data/tiny` as well as a mean stddev file and a vocabulary file. It has to be run for the very first time you run this dataset and is reusable for all further experiments. -- Train your own ASR model - - ```bash - sh run_train.sh - ``` - - `run_train.sh` will start a training job, with training logs printed to stdout and model checkpoint of every pass/epoch saved to `./checkpoints/tiny`. These checkpoints could be used for training resuming, inference, evaluation and deployment. -- Case inference with an existing model - - ```bash - sh run_infer.sh - ``` - - `run_infer.sh` will show us some speech-to-text decoding results for several (default: 10) samples with the trained model. The performance might not be good now as the current model is only trained with a toy subset of LibriSpeech. To see the results with a better model, you can download a well-trained (trained for several days, with the complete LibriSpeech) model and do the inference: - - ```bash - sh run_infer_golden.sh - ``` -- Evaluate an existing model - - ```bash - sh run_test.sh - ``` - - `run_test.sh` will evaluate the model with Word Error Rate (or Character Error Rate) measurement. Similarly, you can also download a well-trained model and test its performance: - - ```bash - sh run_test_golden.sh - ``` - -More detailed information are provided in the following sections. Wish you a happy journey with the *DeepSpeech2 on PaddlePaddle* ASR engine! - - -## Data Preparation - -### Generate Manifest - -*DeepSpeech2 on PaddlePaddle* accepts a textual **manifest** file as its data set interface. A manifest file summarizes a set of speech data, with each line containing some meta data (e.g. filepath, transcription, duration) of one audio clip, in [JSON](http://www.json.org/) format, such as: - -``` -{"audio_filepath": "/home/work/.cache/paddle/Libri/134686/1089-134686-0001.flac", "duration": 3.275, "text": "stuff it into you his belly counselled him"} -{"audio_filepath": "/home/work/.cache/paddle/Libri/134686/1089-134686-0007.flac", "duration": 4.275, "text": "a cold lucid indifference reigned in his soul"} -``` - -To use your custom data, you only need to generate such manifest files to summarize the dataset. Given such summarized manifests, training, inference and all other modules can be aware of where to access the audio files, as well as their meta data including the transcription labels. - -For how to generate such manifest files, please refer to `data/librispeech/librispeech.py`, which will download data and generate manifest files for LibriSpeech dataset. - -### Compute Mean & Stddev for Normalizer - -To perform z-score normalization (zero-mean, unit stddev) upon audio features, we have to estimate in advance the mean and standard deviation of the features, with some training samples: - -```bash -python tools/compute_mean_std.py \ ---num_samples 2000 \ ---specgram_type linear \ ---manifest_paths data/librispeech/manifest.train \ ---output_path data/librispeech/mean_std.npz -``` - -It will compute the mean and standard deviation of power spectrum feature with 2000 random sampled audio clips listed in `data/librispeech/manifest.train` and save the results to `data/librispeech/mean_std.npz` for further usage. - - -### Build Vocabulary - -A vocabulary of possible characters is required to convert the transcription into a list of token indices for training, and in decoding, to convert from a list of indices back to text again. Such a character-based vocabulary can be built with `tools/build_vocab.py`. - -```bash -python tools/build_vocab.py \ ---count_threshold 0 \ ---vocab_path data/librispeech/eng_vocab.txt \ ---manifest_paths data/librispeech/manifest.train -``` - -It will write a vocabuary file `data/librispeeech/eng_vocab.txt` with all transcription text in `data/librispeech/manifest.train`, without vocabulary truncation (`--count_threshold 0`). - -### More Help - -For more help on arguments: - -```bash -python data/librispeech/librispeech.py --help -python tools/compute_mean_std.py --help -python tools/build_vocab.py --help -``` - -## Training a model - -`train.py` is the main caller of the training module. Examples of usage are shown below. - -- Start training from scratch with 8 GPUs: - - ``` - CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 python train.py --trainer_count 8 - ``` - -- Start training from scratch with 16 CPUs: - - ``` - python train.py --use_gpu False --trainer_count 16 - ``` -- Resume training from a checkpoint: - - ``` - CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ - python train.py \ - --init_model_path CHECKPOINT_PATH_TO_RESUME_FROM - ``` - -For more help on arguments: - -```bash -python train.py --help -``` -or refer to `example/librispeech/run_train.sh`. - -## Data Augmentation Pipeline - -Data augmentation has often been a highly effective technique to boost the deep learning performance. We augment our speech data by synthesizing new audios with small random perturbation (label-invariant transformation) added upon raw audios. You don't have to do the syntheses on your own, as it is already embedded into the data provider and is done on the fly, randomly for each epoch during training. - -Six optional augmentation components are provided to be selected, configured and inserted into the processing pipeline. - - - Volume Perturbation - - Speed Perturbation - - Shifting Perturbation - - Online Bayesian normalization - - Noise Perturbation (need background noise audio files) - - Impulse Response (need impulse audio files) - -In order to inform the trainer of what augmentation components are needed and what their processing orders are, it is required to prepare in advance an *augmentation configuration file* in [JSON](http://www.json.org/) format. For example: - -``` -[{ - "type": "speed", - "params": {"min_speed_rate": 0.95, - "max_speed_rate": 1.05}, - "prob": 0.6 -}, -{ - "type": "shift", - "params": {"min_shift_ms": -5, - "max_shift_ms": 5}, - "prob": 0.8 -}] -``` - -When the `--augment_conf_file` argument of `trainer.py` is set to the path of the above example configuration file, every audio clip in every epoch will be processed: with 60% of chance, it will first be speed perturbed with a uniformly random sampled speed-rate between 0.95 and 1.05, and then with 80% of chance it will be shifted in time with a random sampled offset between -5 ms and 5 ms. Finally this newly synthesized audio clip will be feed into the feature extractor for further training. - -For other configuration examples, please refer to `conf/augmenatation.config.example`. - -Be careful when utilizing the data augmentation technique, as improper augmentation will do harm to the training, due to the enlarged train-test gap. - -## Inference and Evaluation - -### Prepare Language Model - -A language model is required to improve the decoder's performance. We have prepared two language models (with lossy compression) for users to download and try. One is for English and the other is for Mandarin. Users can simply run this to download the preprared language models: - -```bash -cd models/lm -sh download_lm_en.sh -sh download_lm_ch.sh -``` - -If you wish to train your own better language model, please refer to [KenLM](https://github.com/kpu/kenlm) for tutorials. Here we provide some tips to show how we preparing our English and Mandarin language models. You can take it as a reference when you train your own. - -#### English LM - -The English corpus is from the [Common Crawl Repository](http://commoncrawl.org) and you can download it from [statmt](http://data.statmt.org/ngrams/deduped_en). We use part en.00 to train our English language model. There are some preprocessing steps before training: - - * Characters not in \[A-Za-z0-9\s'\] (\s represents whitespace characters) are removed and Arabic numbers are converted to English numbers like 1000 to one thousand. - * Repeated whitespace characters are squeezed to one and the beginning whitespace characters are removed. Notice that all transcriptions are lowercase, so all characters are converted to lowercase. - * Top 400,000 most frequent words are selected to build the vocabulary and the rest are replaced with 'UNKNOWNWORD'. - -Now the preprocessing is done and we get a clean corpus to train the language model. Our released language model are trained with agruments '-o 5 --prune 0 1 1 1 1'. '-o 5' means the max order of language model is 5. '--prune 0 1 1 1 1' represents count thresholds for each order and more specifically it will prune singletons for orders two and higher. To save disk storage we convert the arpa file to 'trie' binary file with arguments '-a 22 -q 8 -b 8'. '-a' represents the maximum number of leading bits of pointers in 'trie' to chop. '-q -b' are quantization parameters for probability and backoff. - -#### Mandarin LM - -Different from the English language model, Mandarin language model is character-based where each token is a Chinese character. We use internal corpus to train the released Mandarin language models. The corpus contain billions of tokens. The preprocessing has tiny difference from English language model and main steps include: - - * The beginning and trailing whitespace characters are removed. - * English punctuations and Chinese punctuations are removed. - * A whitespace character between two tokens is inserted. - -Please notice that the released language models only contain Chinese simplified characters. After preprocessing done we can begin to train the language model. The key training arguments for small LM is '-o 5 --prune 0 1 2 4 4' and '-o 5' for large LM. Please refer above section for the meaning of each argument. We also convert the arpa file to binary file using default settings. - -### Speech-to-text Inference - -An inference module caller `infer.py` is provided to infer, decode and visualize speech-to-text results for several given audio clips. It might help to have an intuitive and qualitative evaluation of the ASR model's performance. - -- Inference with GPU: - - ```bash - CUDA_VISIBLE_DEVICES=0 python infer.py --trainer_count 1 - ``` - -- Inference with CPUs: - - ```bash - python infer.py --use_gpu False --trainer_count 12 - ``` - -We provide two types of CTC decoders: *CTC greedy decoder* and *CTC beam search decoder*. The *CTC greedy decoder* is an implementation of the simple best-path decoding algorithm, selecting at each timestep the most likely token, thus being greedy and locally optimal. The [*CTC beam search decoder*](https://arxiv.org/abs/1408.2873) otherwise utilizes a heuristic breadth-first graph search for reaching a near global optimality; it also requires a pre-trained KenLM language model for better scoring and ranking. The decoder type can be set with argument `--decoding_method`. - -For more help on arguments: - -``` -python infer.py --help -``` -or refer to `example/librispeech/run_infer.sh`. - -### Evaluate a Model - -To evaluate a model's performance quantitatively, please run: - -- Evaluation with GPUs: - - ```bash - CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 python test.py --trainer_count 8 - ``` - -- Evaluation with CPUs: - - ```bash - python test.py --use_gpu False --trainer_count 12 - ``` - -The error rate (default: word error rate; can be set with `--error_rate_type`) will be printed. - -For more help on arguments: - -```bash -python test.py --help -``` -or refer to `example/librispeech/run_test.sh`. - -## Hyper-parameters Tuning - -The hyper-parameters $\alpha$ (language model weight) and $\beta$ (word insertion weight) for the [*CTC beam search decoder*](https://arxiv.org/abs/1408.2873) often have a significant impact on the decoder's performance. It would be better to re-tune them on the validation set when the acoustic model is renewed. - -`tools/tune.py` performs a 2-D grid search over the hyper-parameter $\alpha$ and $\beta$. You must provide the range of $\alpha$ and $\beta$, as well as the number of their attempts. - -- Tuning with GPU: - - ```bash - CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ - python tools/tune.py \ - --trainer_count 8 \ - --alpha_from 1.0 \ - --alpha_to 3.2 \ - --num_alphas 45 \ - --beta_from 0.1 \ - --beta_to 0.45 \ - --num_betas 8 - ``` - -- Tuning with CPU: - - ```bash - python tools/tune.py --use_gpu False - ``` - The grid search will print the WER (word error rate) or CER (character error rate) at each point in the hyper-parameters space, and draw the error surface optionally. A proper hyper-parameters range should include the global minima of the error surface for WER/CER, as illustrated in the following figure. - -

- -
An example error surface for tuning on the dev-clean set of LibriSpeech -

- -Usually, as the figure shows, the variation of language model weight ($\alpha$) significantly affect the performance of CTC beam search decoder. And a better procedure is to first tune on serveral data batches (the number can be specified) to find out the proper range of hyper-parameters, then change to the whole validation set to carray out an accurate tuning. - -After tuning, you can reset $\alpha$ and $\beta$ in the inference and evaluation modules to see if they really help improve the ASR performance. For more help - -```bash -python tune.py --help -``` -or refer to `example/librispeech/run_tune.sh`. - -## Running in Docker Container - -Docker is an open source tool to build, ship, and run distributed applications in an isolated environment. A Docker image for this project has been provided in [hub.docker.com](https://hub.docker.com) with all the dependencies installed, including the pre-built PaddlePaddle, CTC decoders, and other necessary Python and third-party packages. This Docker image requires the support of NVIDIA GPU, so please make sure its availiability and the [nvidia-docker](https://github.com/NVIDIA/nvidia-docker) has been installed. - -Take several steps to launch the Docker image: - -- Download the Docker image - -```bash -nvidia-docker pull paddlepaddle/models:deep-speech-2 -``` - -- Clone this repository - -``` -git clone https://github.com/PaddlePaddle/models.git -``` - -- Run the Docker image - -```bash -sudo nvidia-docker run -it -v $(pwd)/models:/models paddlepaddle/models:deep-speech-2 /bin/bash -``` -Now go back and start from the [Getting Started](#getting-started) section, you can execute training, inference and hyper-parameters tuning similarly in the Docker container. - -## Distributed Cloud Training - -We also provide a cloud training module for users to do the distributed cluster training on [PaddleCloud](https://github.com/PaddlePaddle/cloud), to achieve a much faster training speed with multiple machines. To start with this, please first install PaddleCloud client and register a PaddleCloud account, as described in [PaddleCloud Usage](https://github.com/PaddlePaddle/cloud/blob/develop/doc/usage_cn.md#%E4%B8%8B%E8%BD%BD%E5%B9%B6%E9%85%8D%E7%BD%AEpaddlecloud). - -Please take the following steps to submit a training job: - -- Go to directory: - - ```bash - cd cloud - ``` -- Upload data: - - Data must be uploaded to PaddleCloud filesystem to be accessed within a cloud job. `pcloud_upload_data.sh` helps do the data packing and uploading: - - ```bash - sh pcloud_upload_data.sh - ``` - - Given input manifests, `pcloud_upload_data.sh` will: - - - Extract the audio files listed in the input manifests. - - Pack them into a specified number of tar files. - - Upload these tar files to PaddleCloud filesystem. - - Create cloud manifests by replacing local filesystem paths with PaddleCloud filesystem paths. New manifests will be used to inform the cloud jobs of audio files' location and their meta information. - - It should be done only once for the very first time to do the cloud training. Later, the data is kept persisitent on the cloud filesystem and reusable for further job submissions. - - For argument details please refer to [Train DeepSpeech2 on PaddleCloud](https://github.com/PaddlePaddle/models/tree/develop/deep_speech_2/cloud). - - - Configure training arguments: - - Configure the cloud job parameters in `pcloud_submit.sh` (e.g. `NUM_NODES`, `NUM_GPUS`, `CLOUD_TRAIN_DIR`, `JOB_NAME` etc.) and then configure other hyper-parameters for training in `pcloud_train.sh` (just as what you do for local training). - - For argument details please refer to [Train DeepSpeech2 on PaddleCloud](https://github.com/PaddlePaddle/models/tree/develop/deep_speech_2/cloud). - - - Submit the job: - - By running: - - ```bash - sh pcloud_submit.sh - ``` - a training job has been submitted to PaddleCloud, with the job name printed to the console. - - - Get training logs - - Run this to list all the jobs you have submitted, as well as their running status: - - ```bash - paddlecloud get jobs - ``` - - Run this, the corresponding job's logs will be printed. - ```bash - paddlecloud logs -n 10000 $REPLACED_WITH_YOUR_ACTUAL_JOB_NAME - ``` - -For more information about the usage of PaddleCloud, please refer to [PaddleCloud Usage](https://github.com/PaddlePaddle/cloud/blob/develop/doc/usage_cn.md#提交任务). - -For more information about the DeepSpeech2 training on PaddleCloud, please refer to -[Train DeepSpeech2 on PaddleCloud](https://github.com/PaddlePaddle/models/tree/develop/deep_speech_2/cloud). - -## Training for Mandarin Language - -The key steps of training for Mandarin language are same to that of English language and we have also provided an example for Mandarin training with Aishell in ```examples/aishell```. As mentioned above, please execute ```sh run_data.sh```, ```sh run_train.sh```, ```sh run_test.sh``` and ```sh run_infer.sh``` to do data preparation, training, testing and inference correspondingly. We have also prepared a pre-trained model (downloaded by ./models/aishell/download_model.sh) for users to try with ```sh run_infer_golden.sh``` and ```sh run_test_golden.sh```. Notice that, different from English LM, the Mandarin LM is character-based and please run ```tools/tune.py``` to find an optimal setting. - -## Trying Live Demo with Your Own Voice - -Until now, an ASR model is trained and tested qualitatively (`infer.py`) and quantitatively (`test.py`) with existing audio files. But it is not yet tested with your own speech. `deploy/demo_server.py` and `deploy/demo_client.py` helps quickly build up a real-time demo ASR engine with the trained model, enabling you to test and play around with the demo, with your own voice. - -To start the demo's server, please run this in one console: - -```bash -CUDA_VISIBLE_DEVICES=0 \ -python deploy/demo_server.py \ ---trainer_count 1 \ ---host_ip localhost \ ---host_port 8086 -``` - -For the machine (might not be the same machine) to run the demo's client, please do the following installation before moving on. - -For example, on MAC OS X: - -```bash -brew install portaudio -pip install pyaudio -pip install pynput -``` - -Then to start the client, please run this in another console: - -```bash -CUDA_VISIBLE_DEVICES=0 \ -python -u deploy/demo_client.py \ ---host_ip 'localhost' \ ---host_port 8086 -``` - -Now, in the client console, press the `whitespace` key, hold, and start speaking. Until finishing your utterance, release the key to let the speech-to-text results shown in the console. To quit the client, just press `ESC` key. - -Notice that `deploy/demo_client.py` must be run on a machine with a microphone device, while `deploy/demo_server.py` could be run on one without any audio recording hardware, e.g. any remote server machine. Just be careful to set the `host_ip` and `host_port` argument with the actual accessible IP address and port, if the server and client are running with two separate machines. Nothing should be done if they are running on one single machine. - -Please also refer to `examples/mandarin/run_demo_server.sh`, which will first download a pre-trained Mandarin model (trained with 3000 hours of internal speech data) and then start the demo server with the model. With running `examples/mandarin/run_demo_client.sh`, you can speak Mandarin to test it. If you would like to try some other models, just update `--model_path` argument in the script.   - -For more help on arguments: - -```bash -python deploy/demo_server.py --help -python deploy/demo_client.py --help -``` - -## Released Models - -#### Speech Model Released - -Language | Model Name | Training Data | Hours of Speech -:-----------: | :------------: | :----------: | -------: -English | [LibriSpeech Model](http://cloud.dlnel.org/filepub/?uuid=17404caf-cf19-492f-9707-1fad07c19aae) | [LibriSpeech Dataset](http://www.openslr.org/12/) | 960 h -English | [BaiduEN8k Model](to-be-added) | Baidu Internal English Dataset | 8628 h -Mandarin | [Aishell Model](http://cloud.dlnel.org/filepub/?uuid=6c83b9d8-3255-4adf-9726-0fe0be3d0274) | [Aishell Dataset](http://www.openslr.org/33/) | 151 h -Mandarin | [BaiduCN1.2k Model](to-be-added) | Baidu Internal Mandarin Dataset | 1204 h - -#### Language Model Released - -Language Model | Training Data | Token-based | Size | Descriptions -:-------------:| :------------:| :-----: | -----: | :----------------- -[English LM](http://paddlepaddle.bj.bcebos.com/model_zoo/speech/common_crawl_00.prune01111.trie.klm) | [CommonCrawl(en.00)](http://web-language-models.s3-website-us-east-1.amazonaws.com/ngrams/en/deduped/en.00.deduped.xz) | Word-based | 8.3 GB | Pruned with 0 1 1 1 1;
About 1.85 billion n-grams;
'trie' binary with '-a 22 -q 8 -b 8' -[Mandarin LM Small](http://cloud.dlnel.org/filepub/?uuid=d21861e4-4ed6-45bb-ad8e-ae417a43195e) | Baidu Internal Corpus | Char-based | 2.8 GB | Pruned with 0 1 2 4 4;
About 0.13 billion n-grams;
'probing' binary with default settings -[Mandarin LM Large](http://cloud.dlnel.org/filepub/?uuid=245d02bb-cd01-4ebe-b079-b97be864ec37) | Baidu Internal Corpus | Char-based | 70.4 GB | No Pruning;
About 3.7 billion n-grams;
'probing' binary with default settings - -## Experiments and Benchmarks - -#### Benchmark Results for English Models (Word Error Rate) - -Test Set | LibriSpeech Model | BaiduEN8K Model -:--------------------- | ---------------: | -------------------: -LibriSpeech Test-Clean | 7.77 | 6.63 -LibriSpeech Test-Other | 23.25 | 16.59 -VoxForge American-Canadian | 12.52 |   7.46 -VoxForge Commonwealth | 21.08 | 16.23 -VoxForge European | 31.21 | 20.47 -VoxForge Indian | 56.79 | 28.15 -Baidu Internal Testset  |   47.73 |   8.92 - -#### Benchmark Results for Mandarin Model (Character Error Rate) - -Test Set | Aishell Model | BaiduCN1.2k Model -:--------------------- | ---------------: | -------------------: -Baidu Internal Testset | - | 15.49 - -#### Acceleration with Multi-GPUs - -We compare the training time with 1, 2, 4, 8, 16 Tesla K40m GPUs (with a subset of LibriSpeech samples whose audio durations are between 6.0 and 7.0 seconds). And it shows that a **near-linear** acceleration with multiple GPUs has been achieved. In the following figure, the time (in seconds) cost for training is printed on the blue bars. - -
- -| # of GPU | Acceleration Rate | -| -------- | --------------: | -| 1 | 1.00 X | -| 2 | 1.97 X | -| 4 | 3.74 X | -| 8 | 6.21 X | -|16 | 10.70 X | - -`tools/profile.sh` provides such a profiling tool. - -## Questions and Help - -You are welcome to submit questions and bug reports in [Github Issues](https://github.com/PaddlePaddle/models/issues). You are also welcome to contribute to this project. diff --git a/deep_speech_2/cloud/README.md b/deep_speech_2/cloud/README.md deleted file mode 100644 index a5be1c420880d4f32d472cdd23124cbf35033094..0000000000000000000000000000000000000000 --- a/deep_speech_2/cloud/README.md +++ /dev/null @@ -1,63 +0,0 @@ -# Train DeepSpeech2 on PaddleCloud - ->Note: ->Please make sure [PaddleCloud Client](https://github.com/PaddlePaddle/cloud/blob/develop/doc/usage_cn.md#%E4%B8%8B%E8%BD%BD%E5%B9%B6%E9%85%8D%E7%BD%AEpaddlecloud) has be installed and current directory is `deep_speech_2/cloud/` - -## Step 1: Upload Data - -Provided with several input manifests, `pcloud_upload_data.sh` will pack and upload all the containing audio files to PaddleCloud filesystem, and also generate some corresponding manifest files with updated cloud paths. - -Please modify the following arguments in `pcloud_upload_data.sh`: - -- `IN_MANIFESTS`： Paths (in local filesystem) of manifest files containing the audio files to be uploaded. Multiple paths can be concatenated with a whitespace delimeter. -- `OUT_MANIFESTS`: Paths (in local filesystem) to write the updated output manifest files to. Multiple paths can be concatenated with a whitespace delimeter. The values of `audio_filepath` in the output manifests are updated with cloud filesystem paths. -- `CLOUD_DATA_DIR`: Directory (in PaddleCloud filesystem) to upload the data to. Don't forget to replace `USERNAME` in the default directory and make sure that you have the permission to write it. -- `NUM_SHARDS`: Number of data shards / parts (in tar files) to be generated when packing and uploading data. Smaller `num_shards` requires larger temoporal local disk space for packing data. - -By running: - -``` -sh pcloud_upload_data.sh -``` -all the audio files will be uploaded to PaddleCloud filesystem, and you will get modified manifests files in `OUT_MANIFESTS`. - -You have to take this step only once, in the very first time you do the cloud training. Later on, the data is persisitent on the cloud filesystem and reusable for further job submissions. - -## Step 2: Configure Training - -Configure cloud training arguments in `pcloud_submit.sh`, with the following arguments: - -- `TRAIN_MANIFEST`: Manifest filepath (in local filesystem) for training. Notice that the`audio_filepath` should be in cloud filesystem, like those generated by `pcloud_upload_data.sh`. -- `DEV_MANIFEST`: Manifest filepath (in local filesystem) for validation. -- `CLOUD_MODEL_DIR`: Directory (in PaddleCloud filesystem) to save the model parameters (checkpoints). Don't forget to replace `USERNAME` in the default directory and make sure that you have the permission to write it. -- `BATCH_SIZE`: Training batch size for a single node. -- `NUM_GPU`: Number of GPUs allocated for a single node. -- `NUM_NODE`: Number of nodes (machines) allocated for this job. -- `IS_LOCAL`: Set to False to enable parameter server, if using multiple nodes. - -Configure other training hyper-parameters in `pcloud_train.sh` as you wish, just as what you can do in local training. - -By running: - -``` -sh pcloud_submit.sh -``` -you submit a training job to PaddleCloud. And you will see the job name when the submission is done. - - -## Step 3 Get Job Logs - -Run this to list all the jobs you have submitted, as well as their running status: - -``` -paddlecloud get jobs -``` - -Run this, the corresponding job's logs will be printed. -``` -paddlecloud logs -n 10000 $REPLACED_WITH_YOUR_ACTUAL_JOB_NAME -``` - -## More Help - -For more information about the usage of PaddleCloud, please refer to [PaddleCloud Usage](https://github.com/PaddlePaddle/cloud/blob/develop/doc/usage_cn.md#提交任务). diff --git a/deep_speech_2/cloud/_init_paths.py b/deep_speech_2/cloud/_init_paths.py deleted file mode 100644 index 3305d7488ff1cfb03db7175a53f70c1a107fe52e..0000000000000000000000000000000000000000 --- a/deep_speech_2/cloud/_init_paths.py +++ /dev/null @@ -1,17 +0,0 @@ -"""Set up paths for DS2""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os.path -import sys - - -def add_path(path): - if path not in sys.path: - sys.path.insert(0, path) - - -this_dir = os.path.dirname(__file__) -proj_path = os.path.join(this_dir, '..') -add_path(proj_path) diff --git a/deep_speech_2/cloud/pcloud_submit.sh b/deep_speech_2/cloud/pcloud_submit.sh deleted file mode 100644 index 99e458db96b819019628a26f05b3597ea951aeea..0000000000000000000000000000000000000000 --- a/deep_speech_2/cloud/pcloud_submit.sh +++ /dev/null @@ -1,29 +0,0 @@ -#! /usr/bin/env bash - -TRAIN_MANIFEST="cloud/cloud_manifests/cloud.manifest.train" -DEV_MANIFEST="cloud/cloud_manifests/cloud.manifest.dev" -CLOUD_MODEL_DIR="./checkpoints" -BATCH_SIZE=512 -NUM_GPU=8 -NUM_NODE=1 -IS_LOCAL="True" - -JOB_NAME=deepspeech-`date +%Y%m%d%H%M%S` -DS2_PATH=${PWD%/*} -cp -f pcloud_train.sh ${DS2_PATH} - -paddlecloud submit \ --image bootstrapper:5000/paddlepaddle/pcloud_ds2:latest \ --jobname ${JOB_NAME} \ --cpu ${NUM_GPU} \ --gpu ${NUM_GPU} \ --memory 64Gi \ --parallelism ${NUM_NODE} \ --pscpu 1 \ --pservers 1 \ --psmemory 64Gi \ --passes 1 \ --entry "sh pcloud_train.sh ${TRAIN_MANIFEST} ${DEV_MANIFEST} ${CLOUD_MODEL_DIR} ${NUM_GPU} ${BATCH_SIZE} ${IS_LOCAL}" \ -${DS2_PATH} - -rm ${DS2_PATH}/pcloud_train.sh diff --git a/deep_speech_2/cloud/pcloud_train.sh b/deep_speech_2/cloud/pcloud_train.sh deleted file mode 100644 index d0c47dece91c43d0cbfde1f6eb2dcc96fce36391..0000000000000000000000000000000000000000 --- a/deep_speech_2/cloud/pcloud_train.sh +++ /dev/null @@ -1,46 +0,0 @@ -#! /usr/bin/env bash - -TRAIN_MANIFEST=$1 -DEV_MANIFEST=$2 -MODEL_PATH=$3 -NUM_GPU=$4 -BATCH_SIZE=$5 -IS_LOCAL=$6 - -python ./cloud/split_data.py \ ---in_manifest_path=${TRAIN_MANIFEST} \ ---out_manifest_path='/local.manifest.train' - -python ./cloud/split_data.py \ ---in_manifest_path=${DEV_MANIFEST} \ ---out_manifest_path='/local.manifest.dev' - -mkdir ./logs - -python -u train.py \ ---batch_size=${BATCH_SIZE} \ ---trainer_count=${NUM_GPU} \ ---num_passes=200 \ ---num_proc_data=${NUM_GPU} \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---num_iter_print=100 \ ---learning_rate=5e-4 \ ---max_duration=27.0 \ ---min_duration=0.0 \ ---use_sortagrad=True \ ---use_gru=False \ ---use_gpu=True \ ---is_local=${IS_LOCAL} \ ---share_rnn_weights=True \ ---train_manifest='/local.manifest.train' \ ---dev_manifest='/local.manifest.dev' \ ---mean_std_path='data/librispeech/mean_std.npz' \ ---vocab_path='data/librispeech/vocab.txt' \ ---output_model_dir='./checkpoints' \ ---output_model_dir=${MODEL_PATH} \ ---augment_conf_path='conf/augmentation.config' \ ---specgram_type='linear' \ ---shuffle_method='batch_shuffle_clipped' \ -2>&1 | tee ./logs/train.log diff --git a/deep_speech_2/cloud/pcloud_upload_data.sh b/deep_speech_2/cloud/pcloud_upload_data.sh deleted file mode 100644 index 71bb4af19b3b30f6efc31cb9b60f4f3b330b46b9..0000000000000000000000000000000000000000 --- a/deep_speech_2/cloud/pcloud_upload_data.sh +++ /dev/null @@ -1,22 +0,0 @@ -#! /usr/bin/env bash - -mkdir cloud_manifests - -IN_MANIFESTS="../data/librispeech/manifest.train ../data/librispeech/manifest.dev-clean ../data/librispeech/manifest.test-clean" -OUT_MANIFESTS="cloud_manifests/cloud.manifest.train cloud_manifests/cloud.manifest.dev cloud_manifests/cloud.manifest.test" -CLOUD_DATA_DIR="/pfs/dlnel/home/USERNAME/deepspeech2/data/librispeech" -NUM_SHARDS=50 - -python upload_data.py \ ---in_manifest_paths ${IN_MANIFESTS} \ ---out_manifest_paths ${OUT_MANIFESTS} \ ---cloud_data_dir ${CLOUD_DATA_DIR} \ ---num_shards ${NUM_SHARDS} - -if [ $? -ne 0 ] -then - echo "Upload Data Failed!" - exit 1 -fi - -echo "All Done." diff --git a/deep_speech_2/cloud/split_data.py b/deep_speech_2/cloud/split_data.py deleted file mode 100644 index 3496d52bfb5bf6c249c03dfb4df2937625bd55b5..0000000000000000000000000000000000000000 --- a/deep_speech_2/cloud/split_data.py +++ /dev/null @@ -1,41 +0,0 @@ -"""This tool is used for splitting data into each node of -paddlecloud. This script should be called in paddlecloud. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os -import json -import argparse - -parser = argparse.ArgumentParser(description=__doc__) -parser.add_argument( - "--in_manifest_path", - type=str, - required=True, - help="Input manifest path for all nodes.") -parser.add_argument( - "--out_manifest_path", - type=str, - required=True, - help="Output manifest file path for current node.") -args = parser.parse_args() - - -def split_data(in_manifest_path, out_manifest_path): - with open("/trainer_id", "r") as f: - trainer_id = int(f.readline()[:-1]) - with open("/trainer_count", "r") as f: - trainer_count = int(f.readline()[:-1]) - - out_manifest = [] - for index, json_line in enumerate(open(in_manifest_path, 'r')): - if (index % trainer_count) == trainer_id: - out_manifest.append("%s\n" % json_line.strip()) - with open(out_manifest_path, 'w') as f: - f.writelines(out_manifest) - - -if __name__ == '__main__': - split_data(args.in_manifest_path, args.out_manifest_path) diff --git a/deep_speech_2/cloud/upload_data.py b/deep_speech_2/cloud/upload_data.py deleted file mode 100644 index 9973f8c768410fd86a6ded6a74dac24f9f918173..0000000000000000000000000000000000000000 --- a/deep_speech_2/cloud/upload_data.py +++ /dev/null @@ -1,129 +0,0 @@ -"""This script is for uploading data for DeepSpeech2 training on paddlecloud. - -Steps: -1. Read original manifests and extract local sound files. -2. Tar all local sound files into multiple tar files and upload them. -3. Modify original manifests with updated paths in cloud filesystem. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import json -import os -import tarfile -import sys -import argparse -import shutil -from subprocess import call -import _init_paths -from data_utils.utils import read_manifest - -parser = argparse.ArgumentParser(description=__doc__) -parser.add_argument( - "--in_manifest_paths", - default=[ - "../datasets/manifest.train", "../datasets/manifest.dev", - "../datasets/manifest.test" - ], - type=str, - nargs='+', - help="Local filepaths of input manifests to load, pack and upload." - "(default: %(default)s)") -parser.add_argument( - "--out_manifest_paths", - default=[ - "./cloud.manifest.train", "./cloud.manifest.dev", - "./cloud.manifest.test" - ], - type=str, - nargs='+', - help="Local filepaths of modified manifests to write to. " - "(default: %(default)s)") -parser.add_argument( - "--cloud_data_dir", - required=True, - type=str, - help="Destination directory on paddlecloud to upload data to.") -parser.add_argument( - "--num_shards", - default=10, - type=int, - help="Number of parts to split data to. (default: %(default)s)") -parser.add_argument( - "--local_tmp_dir", - default="./tmp/", - type=str, - help="Local directory for storing temporary data. (default: %(default)s)") -args = parser.parse_args() - - -def upload_data(in_manifest_path_list, out_manifest_path_list, local_tmp_dir, - upload_tar_dir, num_shards): - """Extract and pack sound files listed in the manifest files into multple - tar files and upload them to padldecloud. Besides, generate new manifest - files with updated paths in paddlecloud. - """ - # compute total audio number - total_line = 0 - for manifest_path in in_manifest_path_list: - with open(manifest_path, 'r') as f: - total_line += len(f.readlines()) - line_per_tar = (total_line // num_shards) + 1 - - # pack and upload shard by shard - line_count, tar_file = 0, None - for manifest_path, out_manifest_path in zip(in_manifest_path_list, - out_manifest_path_list): - manifest = read_manifest(manifest_path) - out_manifest = [] - for json_data in manifest: - sound_filepath = json_data['audio_filepath'] - sound_filename = os.path.basename(sound_filepath) - if line_count % line_per_tar == 0: - if tar_file != None: - tar_file.close() - pcloud_cp(tar_path, upload_tar_dir) - os.remove(tar_path) - tar_name = 'part-%s-of-%s.tar' % ( - str(line_count // line_per_tar).zfill(5), - str(num_shards).zfill(5)) - tar_path = os.path.join(local_tmp_dir, tar_name) - tar_file = tarfile.open(tar_path, 'w') - tar_file.add(sound_filepath, arcname=sound_filename) - line_count += 1 - json_data['audio_filepath'] = "tar:%s#%s" % ( - os.path.join(upload_tar_dir, tar_name), sound_filename) - out_manifest.append("%s\n" % json.dumps(json_data)) - with open(out_manifest_path, 'w') as f: - f.writelines(out_manifest) - pcloud_cp(out_manifest_path, upload_tar_dir) - tar_file.close() - pcloud_cp(tar_path, upload_tar_dir) - os.remove(tar_path) - - -def pcloud_mkdir(dir): - """Make directory in PaddleCloud filesystem. - """ - if call(['paddlecloud', 'mkdir', dir]) != 0: - raise IOError("PaddleCloud mkdir failed: %s." % dir) - - -def pcloud_cp(src, dst): - """Copy src from local filesytem to dst in PaddleCloud filesystem, - or downlowd src from PaddleCloud filesystem to dst in local filesystem. - """ - if call(['paddlecloud', 'cp', src, dst]) != 0: - raise IOError("PaddleCloud cp failed: from [%s] to [%s]." % (src, dst)) - - -if __name__ == '__main__': - if not os.path.exists(args.local_tmp_dir): - os.makedirs(args.local_tmp_dir) - pcloud_mkdir(args.cloud_data_dir) - - upload_data(args.in_manifest_paths, args.out_manifest_paths, - args.local_tmp_dir, args.cloud_data_dir, args.num_shards) - - shutil.rmtree(args.local_tmp_dir) diff --git a/deep_speech_2/conf/augmentation.config b/deep_speech_2/conf/augmentation.config deleted file mode 100644 index 6c24da5497460d4bae9c9c4fecdbe96ab8da7532..0000000000000000000000000000000000000000 --- a/deep_speech_2/conf/augmentation.config +++ /dev/null @@ -1,8 +0,0 @@ -[ - { - "type": "shift", - "params": {"min_shift_ms": -5, - "max_shift_ms": 5}, - "prob": 1.0 - } -] diff --git a/deep_speech_2/conf/augmentation.config.example b/deep_speech_2/conf/augmentation.config.example deleted file mode 100644 index 21ed6ee10375a749f4c072389509db2020d9e9c9..0000000000000000000000000000000000000000 --- a/deep_speech_2/conf/augmentation.config.example +++ /dev/null @@ -1,39 +0,0 @@ -[ - { - "type": "noise", - "params": {"min_snr_dB": 40, - "max_snr_dB": 50, - "noise_manifest_path": "datasets/manifest.noise"}, - "prob": 0.6 - }, - { - "type": "impulse", - "params": {"impulse_manifest_path": "datasets/manifest.impulse"}, - "prob": 0.5 - }, - { - "type": "speed", - "params": {"min_speed_rate": 0.95, - "max_speed_rate": 1.05}, - "prob": 0.5 - }, - { - "type": "shift", - "params": {"min_shift_ms": -5, - "max_shift_ms": 5}, - "prob": 1.0 - }, - { - "type": "volume", - "params": {"min_gain_dBFS": -10, - "max_gain_dBFS": 10}, - "prob": 0.0 - }, - { - "type": "bayesian_normal", - "params": {"target_db": -20, - "prior_db": -20, - "prior_samples": 100}, - "prob": 0.0 - } -] diff --git a/deep_speech_2/data/aishell/aishell.py b/deep_speech_2/data/aishell/aishell.py deleted file mode 100644 index 17786b5d42d19fd1300c142b494d78f56e9f26dd..0000000000000000000000000000000000000000 --- a/deep_speech_2/data/aishell/aishell.py +++ /dev/null @@ -1,109 +0,0 @@ -"""Prepare Aishell mandarin dataset - -Download, unpack and create manifest files. -Manifest file is a json-format file with each line containing the -meta data (i.e. audio filepath, transcript and audio duration) -of each audio file in the data set. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os -import codecs -import soundfile -import json -import argparse -from data_utils.utility import download, unpack - -DATA_HOME = os.path.expanduser('~/.cache/paddle/dataset/speech') - -URL_ROOT = 'http://www.openslr.org/resources/33' -DATA_URL = URL_ROOT + '/data_aishell.tgz' -MD5_DATA = '2f494334227864a8a8fec932999db9d8' - -parser = argparse.ArgumentParser(description=__doc__) -parser.add_argument( - "--target_dir", - default=DATA_HOME + "/Aishell", - type=str, - help="Directory to save the dataset. (default: %(default)s)") -parser.add_argument( - "--manifest_prefix", - default="manifest", - type=str, - help="Filepath prefix for output manifests. (default: %(default)s)") -args = parser.parse_args() - - -def create_manifest(data_dir, manifest_path_prefix): - print("Creating manifest %s ..." % manifest_path_prefix) - json_lines = [] - transcript_path = os.path.join(data_dir, 'transcript', - 'aishell_transcript_v0.8.txt') - transcript_dict = {} - for line in codecs.open(transcript_path, 'r', 'utf-8'): - line = line.strip() - if line == '': continue - audio_id, text = line.split(' ', 1) - # remove withespace - text = ''.join(text.split()) - transcript_dict[audio_id] = text - - data_types = ['train', 'dev', 'test'] - for type in data_types: - audio_dir = os.path.join(data_dir, 'wav', type) - for subfolder, _, filelist in sorted(os.walk(audio_dir)): - for fname in filelist: - audio_path = os.path.join(subfolder, fname) - audio_id = fname[:-4] - # if no transcription for audio then skipped - if audio_id not in transcript_dict: - continue - audio_data, samplerate = soundfile.read(audio_path) - duration = float(len(audio_data) / samplerate) - text = transcript_dict[audio_id] - json_lines.append( - json.dumps( - { - 'audio_filepath': audio_path, - 'duration': duration, - 'text': text - }, - ensure_ascii=False)) - manifest_path = manifest_path_prefix + '.' + type - with codecs.open(manifest_path, 'w', 'utf-8') as fout: - for line in json_lines: - fout.write(line + '\n') - - -def prepare_dataset(url, md5sum, target_dir, manifest_path): - """Download, unpack and create manifest file.""" - data_dir = os.path.join(target_dir, 'data_aishell') - if not os.path.exists(data_dir): - filepath = download(url, md5sum, target_dir) - unpack(filepath, target_dir) - # unpack all audio tar files - audio_dir = os.path.join(data_dir, 'wav') - for subfolder, _, filelist in sorted(os.walk(audio_dir)): - for ftar in filelist: - unpack(os.path.join(subfolder, ftar), subfolder, True) - else: - print("Skip downloading and unpacking. Data already exists in %s." % - target_dir) - create_manifest(data_dir, manifest_path) - - -def main(): - if args.target_dir.startswith('~'): - args.target_dir = os.path.expanduser(args.target_dir) - - prepare_dataset( - url=DATA_URL, - md5sum=MD5_DATA, - target_dir=args.target_dir, - manifest_path=args.manifest_prefix) - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/data/librispeech/librispeech.py b/deep_speech_2/data/librispeech/librispeech.py deleted file mode 100644 index 9a8e1c2871f74823b04c5839dd43f08f9a03d1df..0000000000000000000000000000000000000000 --- a/deep_speech_2/data/librispeech/librispeech.py +++ /dev/null @@ -1,148 +0,0 @@ -"""Prepare Librispeech ASR datasets. - -Download, unpack and create manifest files. -Manifest file is a json-format file with each line containing the -meta data (i.e. audio filepath, transcript and audio duration) -of each audio file in the data set. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import distutils.util -import os -import sys -import argparse -import soundfile -import json -import codecs -from data_utils.utility import download, unpack - -URL_ROOT = "http://www.openslr.org/resources/12" -URL_TEST_CLEAN = URL_ROOT + "/test-clean.tar.gz" -URL_TEST_OTHER = URL_ROOT + "/test-other.tar.gz" -URL_DEV_CLEAN = URL_ROOT + "/dev-clean.tar.gz" -URL_DEV_OTHER = URL_ROOT + "/dev-other.tar.gz" -URL_TRAIN_CLEAN_100 = URL_ROOT + "/train-clean-100.tar.gz" -URL_TRAIN_CLEAN_360 = URL_ROOT + "/train-clean-360.tar.gz" -URL_TRAIN_OTHER_500 = URL_ROOT + "/train-other-500.tar.gz" - -MD5_TEST_CLEAN = "32fa31d27d2e1cad72775fee3f4849a9" -MD5_TEST_OTHER = "fb5a50374b501bb3bac4815ee91d3135" -MD5_DEV_CLEAN = "42e2234ba48799c1f50f24a7926300a1" -MD5_DEV_OTHER = "c8d0bcc9cca99d4f8b62fcc847357931" -MD5_TRAIN_CLEAN_100 = "2a93770f6d5c6c964bc36631d331a522" -MD5_TRAIN_CLEAN_360 = "c0e676e450a7ff2f54aeade5171606fa" -MD5_TRAIN_OTHER_500 = "d1a0fd59409feb2c614ce4d30c387708" - -parser = argparse.ArgumentParser(description=__doc__) -parser.add_argument( - "--target_dir", - default='~/.cache/paddle/dataset/speech/libri', - type=str, - help="Directory to save the dataset. (default: %(default)s)") -parser.add_argument( - "--manifest_prefix", - default="manifest", - type=str, - help="Filepath prefix for output manifests. (default: %(default)s)") -parser.add_argument( - "--full_download", - default="True", - type=distutils.util.strtobool, - help="Download all datasets for Librispeech." - " If False, only download a minimal requirement (test-clean, dev-clean" - " train-clean-100). (default: %(default)s)") -args = parser.parse_args() - - -def create_manifest(data_dir, manifest_path): - """Create a manifest json file summarizing the data set, with each line - containing the meta data (i.e. audio filepath, transcription text, audio - duration) of each audio file within the data set. - """ - print("Creating manifest %s ..." % manifest_path) - json_lines = [] - for subfolder, _, filelist in sorted(os.walk(data_dir)): - text_filelist = [ - filename for filename in filelist if filename.endswith('trans.txt') - ] - if len(text_filelist) > 0: - text_filepath = os.path.join(data_dir, subfolder, text_filelist[0]) - for line in open(text_filepath): - segments = line.strip().split() - text = ' '.join(segments[1:]).lower() - audio_filepath = os.path.join(data_dir, subfolder, - segments[0] + '.flac') - audio_data, samplerate = soundfile.read(audio_filepath) - duration = float(len(audio_data)) / samplerate - json_lines.append( - json.dumps({ - 'audio_filepath': audio_filepath, - 'duration': duration, - 'text': text - })) - with codecs.open(manifest_path, 'w', 'utf-8') as out_file: - for line in json_lines: - out_file.write(line + '\n') - - -def prepare_dataset(url, md5sum, target_dir, manifest_path): - """Download, unpack and create summmary manifest file. - """ - if not os.path.exists(os.path.join(target_dir, "LibriSpeech")): - # download - filepath = download(url, md5sum, target_dir) - # unpack - unpack(filepath, target_dir) - else: - print("Skip downloading and unpacking. Data already exists in %s." % - target_dir) - # create manifest json file - create_manifest(target_dir, manifest_path) - - -def main(): - if args.target_dir.startswith('~'): - args.target_dir = os.path.expanduser(args.target_dir) - - prepare_dataset( - url=URL_TEST_CLEAN, - md5sum=MD5_TEST_CLEAN, - target_dir=os.path.join(args.target_dir, "test-clean"), - manifest_path=args.manifest_prefix + ".test-clean") - prepare_dataset( - url=URL_DEV_CLEAN, - md5sum=MD5_DEV_CLEAN, - target_dir=os.path.join(args.target_dir, "dev-clean"), - manifest_path=args.manifest_prefix + ".dev-clean") - if args.full_download: - prepare_dataset( - url=URL_TRAIN_CLEAN_100, - md5sum=MD5_TRAIN_CLEAN_100, - target_dir=os.path.join(args.target_dir, "train-clean-100"), - manifest_path=args.manifest_prefix + ".train-clean-100") - prepare_dataset( - url=URL_TEST_OTHER, - md5sum=MD5_TEST_OTHER, - target_dir=os.path.join(args.target_dir, "test-other"), - manifest_path=args.manifest_prefix + ".test-other") - prepare_dataset( - url=URL_DEV_OTHER, - md5sum=MD5_DEV_OTHER, - target_dir=os.path.join(args.target_dir, "dev-other"), - manifest_path=args.manifest_prefix + ".dev-other") - prepare_dataset( - url=URL_TRAIN_CLEAN_360, - md5sum=MD5_TRAIN_CLEAN_360, - target_dir=os.path.join(args.target_dir, "train-clean-360"), - manifest_path=args.manifest_prefix + ".train-clean-360") - prepare_dataset( - url=URL_TRAIN_OTHER_500, - md5sum=MD5_TRAIN_OTHER_500, - target_dir=os.path.join(args.target_dir, "train-other-500"), - manifest_path=args.manifest_prefix + ".train-other-500") - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/data/noise/chime3_background.py b/deep_speech_2/data/noise/chime3_background.py deleted file mode 100644 index f79ca7335bda7aec795bc43c32a51519f3363d85..0000000000000000000000000000000000000000 --- a/deep_speech_2/data/noise/chime3_background.py +++ /dev/null @@ -1,128 +0,0 @@ -"""Prepare CHiME3 background data. - -Download, unpack and create manifest files. -Manifest file is a json-format file with each line containing the -meta data (i.e. audio filepath, transcript and audio duration) -of each audio file in the data set. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import distutils.util -import os -import wget -import zipfile -import argparse -import soundfile -import json -from paddle.v2.dataset.common import md5file - -DATA_HOME = os.path.expanduser('~/.cache/paddle/dataset/speech') - -URL = "https://d4s.myairbridge.com/packagev2/AG0Y3DNBE5IWRRTV/?dlid=W19XG7T0NNHB027139H0EQ" -MD5 = "c3ff512618d7a67d4f85566ea1bc39ec" - -parser = argparse.ArgumentParser(description=__doc__) -parser.add_argument( - "--target_dir", - default=DATA_HOME + "/chime3_background", - type=str, - help="Directory to save the dataset. (default: %(default)s)") -parser.add_argument( - "--manifest_filepath", - default="manifest.chime3.background", - type=str, - help="Filepath for output manifests. (default: %(default)s)") -args = parser.parse_args() - - -def download(url, md5sum, target_dir, filename=None): - """Download file from url to target_dir, and check md5sum.""" - if filename == None: - filename = url.split("/")[-1] - if not os.path.exists(target_dir): os.makedirs(target_dir) - filepath = os.path.join(target_dir, filename) - if not (os.path.exists(filepath) and md5file(filepath) == md5sum): - print("Downloading %s ..." % url) - wget.download(url, target_dir) - print("\nMD5 Chesksum %s ..." % filepath) - if not md5file(filepath) == md5sum: - raise RuntimeError("MD5 checksum failed.") - else: - print("File exists, skip downloading. (%s)" % filepath) - return filepath - - -def unpack(filepath, target_dir): - """Unpack the file to the target_dir.""" - print("Unpacking %s ..." % filepath) - if filepath.endswith('.zip'): - zip = zipfile.ZipFile(filepath, 'r') - zip.extractall(target_dir) - zip.close() - elif filepath.endswith('.tar') or filepath.endswith('.tar.gz'): - tar = zipfile.open(filepath) - tar.extractall(target_dir) - tar.close() - else: - raise ValueError("File format is not supported for unpacking.") - - -def create_manifest(data_dir, manifest_path): - """Create a manifest json file summarizing the data set, with each line - containing the meta data (i.e. audio filepath, transcription text, audio - duration) of each audio file within the data set. - """ - print("Creating manifest %s ..." % manifest_path) - json_lines = [] - for subfolder, _, filelist in sorted(os.walk(data_dir)): - for filename in filelist: - if filename.endswith('.wav'): - filepath = os.path.join(data_dir, subfolder, filename) - audio_data, samplerate = soundfile.read(filepath) - duration = float(len(audio_data)) / samplerate - json_lines.append( - json.dumps({ - 'audio_filepath': filepath, - 'duration': duration, - 'text': '' - })) - with open(manifest_path, 'w') as out_file: - for line in json_lines: - out_file.write(line + '\n') - - -def prepare_chime3(url, md5sum, target_dir, manifest_path): - """Download, unpack and create summmary manifest file.""" - if not os.path.exists(os.path.join(target_dir, "CHiME3")): - # download - filepath = download(url, md5sum, target_dir, - "myairbridge-AG0Y3DNBE5IWRRTV.zip") - # unpack - unpack(filepath, target_dir) - unpack( - os.path.join(target_dir, 'CHiME3_background_bus.zip'), target_dir) - unpack( - os.path.join(target_dir, 'CHiME3_background_caf.zip'), target_dir) - unpack( - os.path.join(target_dir, 'CHiME3_background_ped.zip'), target_dir) - unpack( - os.path.join(target_dir, 'CHiME3_background_str.zip'), target_dir) - else: - print("Skip downloading and unpacking. Data already exists in %s." % - target_dir) - # create manifest json file - create_manifest(target_dir, manifest_path) - - -def main(): - prepare_chime3( - url=URL, - md5sum=MD5, - target_dir=args.target_dir, - manifest_path=args.manifest_filepath) - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/data_utils/audio.py b/deep_speech_2/data_utils/audio.py deleted file mode 100644 index 3fb782951699bc1abd3d5613621e89685ee387de..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/audio.py +++ /dev/null @@ -1,685 +0,0 @@ -"""Contains the audio segment class.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import numpy as np -import io -import struct -import re -import soundfile -import resampy -from scipy import signal -import random -import copy - - -class AudioSegment(object): - """Monaural audio segment abstraction. - - :param samples: Audio samples [num_samples x num_channels]. - :type samples: ndarray.float32 - :param sample_rate: Audio sample rate. - :type sample_rate: int - :raises TypeError: If the sample data type is not float or int. - """ - - def __init__(self, samples, sample_rate): - """Create audio segment from samples. - - Samples are convert float32 internally, with int scaled to [-1, 1]. - """ - self._samples = self._convert_samples_to_float32(samples) - self._sample_rate = sample_rate - if self._samples.ndim >= 2: - self._samples = np.mean(self._samples, 1) - - def __eq__(self, other): - """Return whether two objects are equal.""" - if type(other) is not type(self): - return False - if self._sample_rate != other._sample_rate: - return False - if self._samples.shape != other._samples.shape: - return False - if np.any(self.samples != other._samples): - return False - return True - - def __ne__(self, other): - """Return whether two objects are unequal.""" - return not self.__eq__(other) - - def __str__(self): - """Return human-readable representation of segment.""" - return ("%s: num_samples=%d, sample_rate=%d, duration=%.2fsec, " - "rms=%.2fdB" % (type(self), self.num_samples, self.sample_rate, - self.duration, self.rms_db)) - - @classmethod - def from_file(cls, file): - """Create audio segment from audio file. - - :param filepath: Filepath or file object to audio file. - :type filepath: basestring|file - :return: Audio segment instance. - :rtype: AudioSegment - """ - if isinstance(file, basestring) and re.findall(r".seqbin_\d+$", file): - return cls.from_sequence_file(file) - else: - samples, sample_rate = soundfile.read(file, dtype='float32') - return cls(samples, sample_rate) - - @classmethod - def slice_from_file(cls, file, start=None, end=None): - """Loads a small section of an audio without having to load - the entire file into the memory which can be incredibly wasteful. - - :param file: Input audio filepath or file object. - :type file: basestring|file - :param start: Start time in seconds. If start is negative, it wraps - around from the end. If not provided, this function - reads from the very beginning. - :type start: float - :param end: End time in seconds. If end is negative, it wraps around - from the end. If not provided, the default behvaior is - to read to the end of the file. - :type end: float - :return: AudioSegment instance of the specified slice of the input - audio file. - :rtype: AudioSegment - :raise ValueError: If start or end is incorrectly set, e.g. out of - bounds in time. - """ - sndfile = soundfile.SoundFile(file) - sample_rate = sndfile.samplerate - duration = float(len(sndfile)) / sample_rate - start = 0. if start is None else start - end = 0. if end is None else end - if start < 0.0: - start += duration - if end < 0.0: - end += duration - if start < 0.0: - raise ValueError("The slice start position (%f s) is out of " - "bounds." % start) - if end < 0.0: - raise ValueError("The slice end position (%f s) is out of bounds." % - end) - if start > end: - raise ValueError("The slice start position (%f s) is later than " - "the slice end position (%f s)." % (start, end)) - if end > duration: - raise ValueError("The slice end position (%f s) is out of bounds " - "(> %f s)" % (end, duration)) - start_frame = int(start * sample_rate) - end_frame = int(end * sample_rate) - sndfile.seek(start_frame) - data = sndfile.read(frames=end_frame - start_frame, dtype='float32') - return cls(data, sample_rate) - - @classmethod - def from_sequence_file(cls, filepath): - """Create audio segment from sequence file. Sequence file is a binary - file containing a collection of multiple audio files, with several - header bytes in the head indicating the offsets of each audio byte data - chunk. - - The format is: - - 4 bytes (int, version), - 4 bytes (int, num of utterance), - 4 bytes (int, bytes per header), - [bytes_per_header*(num_utterance+1)] bytes (offsets for each audio), - audio_bytes_data_of_1st_utterance, - audio_bytes_data_of_2nd_utterance, - ...... - - Sequence file name must end with ".seqbin". And the filename of the 5th - utterance's audio file in sequence file "xxx.seqbin" must be - "xxx.seqbin_5", with "5" indicating the utterance index within this - sequence file (starting from 1). - - :param filepath: Filepath of sequence file. - :type filepath: basestring - :return: Audio segment instance. - :rtype: AudioSegment - """ - # parse filepath - matches = re.match(r"(.+\.seqbin)_(\d+)", filepath) - if matches is None: - raise IOError("File type of %s is not supported" % filepath) - filename = matches.group(1) - fileno = int(matches.group(2)) - - # read headers - f = open(filename, 'rb') - version = f.read(4) - num_utterances = struct.unpack("i", f.read(4))[0] - bytes_per_header = struct.unpack("i", f.read(4))[0] - header_bytes = f.read(bytes_per_header * (num_utterances + 1)) - header = [ - struct.unpack("i", header_bytes[bytes_per_header * i: - bytes_per_header * (i + 1)])[0] - for i in range(num_utterances + 1) - ] - - # read audio bytes - f.seek(header[fileno - 1]) - audio_bytes = f.read(header[fileno] - header[fileno - 1]) - f.close() - - # create audio segment - try: - return cls.from_bytes(audio_bytes) - except Exception as e: - samples = np.frombuffer(audio_bytes, dtype='int16') - return cls(samples=samples, sample_rate=8000) - - @classmethod - def from_bytes(cls, bytes): - """Create audio segment from a byte string containing audio samples. - - :param bytes: Byte string containing audio samples. - :type bytes: str - :return: Audio segment instance. - :rtype: AudioSegment - """ - samples, sample_rate = soundfile.read( - io.BytesIO(bytes), dtype='float32') - return cls(samples, sample_rate) - - @classmethod - def concatenate(cls, *segments): - """Concatenate an arbitrary number of audio segments together. - - :param *segments: Input audio segments to be concatenated. - :type *segments: tuple of AudioSegment - :return: Audio segment instance as concatenating results. - :rtype: AudioSegment - :raises ValueError: If the number of segments is zero, or if the - sample_rate of any segments does not match. - :raises TypeError: If any segment is not AudioSegment instance. - """ - # Perform basic sanity-checks. - if len(segments) == 0: - raise ValueError("No audio segments are given to concatenate.") - sample_rate = segments[0]._sample_rate - for seg in segments: - if sample_rate != seg._sample_rate: - raise ValueError("Can't concatenate segments with " - "different sample rates") - if type(seg) is not cls: - raise TypeError("Only audio segments of the same type " - "can be concatenated.") - samples = np.concatenate([seg.samples for seg in segments]) - return cls(samples, sample_rate) - - @classmethod - def make_silence(cls, duration, sample_rate): - """Creates a silent audio segment of the given duration and sample rate. - - :param duration: Length of silence in seconds. - :type duration: float - :param sample_rate: Sample rate. - :type sample_rate: float - :return: Silent AudioSegment instance of the given duration. - :rtype: AudioSegment - """ - samples = np.zeros(int(duration * sample_rate)) - return cls(samples, sample_rate) - - def to_wav_file(self, filepath, dtype='float32'): - """Save audio segment to disk as wav file. - - :param filepath: WAV filepath or file object to save the - audio segment. - :type filepath: basestring|file - :param dtype: Subtype for audio file. Options: 'int16', 'int32', - 'float32', 'float64'. Default is 'float32'. - :type dtype: str - :raises TypeError: If dtype is not supported. - """ - samples = self._convert_samples_from_float32(self._samples, dtype) - subtype_map = { - 'int16': 'PCM_16', - 'int32': 'PCM_32', - 'float32': 'FLOAT', - 'float64': 'DOUBLE' - } - soundfile.write( - filepath, - samples, - self._sample_rate, - format='WAV', - subtype=subtype_map[dtype]) - - def superimpose(self, other): - """Add samples from another segment to those of this segment - (sample-wise addition, not segment concatenation). - - Note that this is an in-place transformation. - - :param other: Segment containing samples to be added in. - :type other: AudioSegments - :raise TypeError: If type of two segments don't match. - :raise ValueError: If the sample rates of the two segments are not - equal, or if the lengths of segments don't match. - """ - if isinstance(other, type(self)): - raise TypeError("Cannot add segments of different types: %s " - "and %s." % (type(self), type(other))) - if self._sample_rate != other._sample_rate: - raise ValueError("Sample rates must match to add segments.") - if len(self._samples) != len(other._samples): - raise ValueError("Segment lengths must match to add segments.") - self._samples += other._samples - - def to_bytes(self, dtype='float32'): - """Create a byte string containing the audio content. - - :param dtype: Data type for export samples. Options: 'int16', 'int32', - 'float32', 'float64'. Default is 'float32'. - :type dtype: str - :return: Byte string containing audio content. - :rtype: str - """ - samples = self._convert_samples_from_float32(self._samples, dtype) - return samples.tostring() - - def gain_db(self, gain): - """Apply gain in decibels to samples. - - Note that this is an in-place transformation. - - :param gain: Gain in decibels to apply to samples. - :type gain: float|1darray - """ - self._samples *= 10.**(gain / 20.) - - def change_speed(self, speed_rate): - """Change the audio speed by linear interpolation. - - Note that this is an in-place transformation. - - :param speed_rate: Rate of speed change: - speed_rate > 1.0, speed up the audio; - speed_rate = 1.0, unchanged; - speed_rate < 1.0, slow down the audio; - speed_rate <= 0.0, not allowed, raise ValueError. - :type speed_rate: float - :raises ValueError: If speed_rate <= 0.0. - """ - if speed_rate <= 0: - raise ValueError("speed_rate should be greater than zero.") - old_length = self._samples.shape[0] - new_length = int(old_length / speed_rate) - old_indices = np.arange(old_length) - new_indices = np.linspace(start=0, stop=old_length, num=new_length) - self._samples = np.interp(new_indices, old_indices, self._samples) - - def normalize(self, target_db=-20, max_gain_db=300.0): - """Normalize audio to be of the desired RMS value in decibels. - - Note that this is an in-place transformation. - - :param target_db: Target RMS value in decibels. This value should be - less than 0.0 as 0.0 is full-scale audio. - :type target_db: float - :param max_gain_db: Max amount of gain in dB that can be applied for - normalization. This is to prevent nans when - attempting to normalize a signal consisting of - all zeros. - :type max_gain_db: float - :raises ValueError: If the required gain to normalize the segment to - the target_db value exceeds max_gain_db. - """ - gain = target_db - self.rms_db - if gain > max_gain_db: - raise ValueError( - "Unable to normalize segment to %f dB because the " - "the probable gain have exceeds max_gain_db (%f dB)" % - (target_db, max_gain_db)) - self.gain_db(min(max_gain_db, target_db - self.rms_db)) - - def normalize_online_bayesian(self, - target_db, - prior_db, - prior_samples, - startup_delay=0.0): - """Normalize audio using a production-compatible online/causal - algorithm. This uses an exponential likelihood and gamma prior to - make online estimates of the RMS even when there are very few samples. - - Note that this is an in-place transformation. - - :param target_db: Target RMS value in decibels. - :type target_bd: float - :param prior_db: Prior RMS estimate in decibels. - :type prior_db: float - :param prior_samples: Prior strength in number of samples. - :type prior_samples: float - :param startup_delay: Default 0.0s. If provided, this function will - accrue statistics for the first startup_delay - seconds before applying online normalization. - :type startup_delay: float - """ - # Estimate total RMS online. - startup_sample_idx = min(self.num_samples - 1, - int(self.sample_rate * startup_delay)) - prior_mean_squared = 10.**(prior_db / 10.) - prior_sum_of_squares = prior_mean_squared * prior_samples - cumsum_of_squares = np.cumsum(self.samples**2) - sample_count = np.arange(self.num_samples) + 1 - if startup_sample_idx > 0: - cumsum_of_squares[:startup_sample_idx] = \ - cumsum_of_squares[startup_sample_idx] - sample_count[:startup_sample_idx] = \ - sample_count[startup_sample_idx] - mean_squared_estimate = ((cumsum_of_squares + prior_sum_of_squares) / - (sample_count + prior_samples)) - rms_estimate_db = 10 * np.log10(mean_squared_estimate) - # Compute required time-varying gain. - gain_db = target_db - rms_estimate_db - self.gain_db(gain_db) - - def resample(self, target_sample_rate, filter='kaiser_best'): - """Resample the audio to a target sample rate. - - Note that this is an in-place transformation. - - :param target_sample_rate: Target sample rate. - :type target_sample_rate: int - :param filter: The resampling filter to use one of {'kaiser_best', - 'kaiser_fast'}. - :type filter: str - """ - self._samples = resampy.resample( - self.samples, self.sample_rate, target_sample_rate, filter=filter) - self._sample_rate = target_sample_rate - - def pad_silence(self, duration, sides='both'): - """Pad this audio sample with a period of silence. - - Note that this is an in-place transformation. - - :param duration: Length of silence in seconds to pad. - :type duration: float - :param sides: Position for padding: - 'beginning' - adds silence in the beginning; - 'end' - adds silence in the end; - 'both' - adds silence in both the beginning and the end. - :type sides: str - :raises ValueError: If sides is not supported. - """ - if duration == 0.0: - return self - cls = type(self) - silence = self.make_silence(duration, self._sample_rate) - if sides == "beginning": - padded = cls.concatenate(silence, self) - elif sides == "end": - padded = cls.concatenate(self, silence) - elif sides == "both": - padded = cls.concatenate(silence, self, silence) - else: - raise ValueError("Unknown value for the sides %s" % sides) - self._samples = padded._samples - - def shift(self, shift_ms): - """Shift the audio in time. If `shift_ms` is positive, shift with time - advance; if negative, shift with time delay. Silence are padded to - keep the duration unchanged. - - Note that this is an in-place transformation. - - :param shift_ms: Shift time in millseconds. If positive, shift with - time advance; if negative; shift with time delay. - :type shift_ms: float - :raises ValueError: If shift_ms is longer than audio duration. - """ - if abs(shift_ms) / 1000.0 > self.duration: - raise ValueError("Absolute value of shift_ms should be smaller " - "than audio duration.") - shift_samples = int(shift_ms * self._sample_rate / 1000) - if shift_samples > 0: - # time advance - self._samples[:-shift_samples] = self._samples[shift_samples:] - self._samples[-shift_samples:] = 0 - elif shift_samples < 0: - # time delay - self._samples[-shift_samples:] = self._samples[:shift_samples] - self._samples[:-shift_samples] = 0 - - def subsegment(self, start_sec=None, end_sec=None): - """Cut the AudioSegment between given boundaries. - - Note that this is an in-place transformation. - - :param start_sec: Beginning of subsegment in seconds. - :type start_sec: float - :param end_sec: End of subsegment in seconds. - :type end_sec: float - :raise ValueError: If start_sec or end_sec is incorrectly set, e.g. out - of bounds in time. - """ - start_sec = 0.0 if start_sec is None else start_sec - end_sec = self.duration if end_sec is None else end_sec - if start_sec < 0.0: - start_sec = self.duration + start_sec - if end_sec < 0.0: - end_sec = self.duration + end_sec - if start_sec < 0.0: - raise ValueError("The slice start position (%f s) is out of " - "bounds." % start_sec) - if end_sec < 0.0: - raise ValueError("The slice end position (%f s) is out of bounds." % - end_sec) - if start_sec > end_sec: - raise ValueError("The slice start position (%f s) is later than " - "the end position (%f s)." % (start_sec, end_sec)) - if end_sec > self.duration: - raise ValueError("The slice end position (%f s) is out of bounds " - "(> %f s)" % (end_sec, self.duration)) - start_sample = int(round(start_sec * self._sample_rate)) - end_sample = int(round(end_sec * self._sample_rate)) - self._samples = self._samples[start_sample:end_sample] - - def random_subsegment(self, subsegment_length, rng=None): - """Cut the specified length of the audiosegment randomly. - - Note that this is an in-place transformation. - - :param subsegment_length: Subsegment length in seconds. - :type subsegment_length: float - :param rng: Random number generator state. - :type rng: random.Random - :raises ValueError: If the length of subsegment is greater than - the origineal segemnt. - """ - rng = random.Random() if rng is None else rng - if subsegment_length > self.duration: - raise ValueError("Length of subsegment must not be greater " - "than original segment.") - start_time = rng.uniform(0.0, self.duration - subsegment_length) - self.subsegment(start_time, start_time + subsegment_length) - - def convolve(self, impulse_segment, allow_resample=False): - """Convolve this audio segment with the given impulse segment. - - Note that this is an in-place transformation. - - :param impulse_segment: Impulse response segments. - :type impulse_segment: AudioSegment - :param allow_resample: Indicates whether resampling is allowed when - the impulse_segment has a different sample - rate from this signal. - :type allow_resample: bool - :raises ValueError: If the sample rate is not match between two - audio segments when resample is not allowed. - """ - if allow_resample and self.sample_rate != impulse_segment.sample_rate: - impulse_segment.resample(self.sample_rate) - if self.sample_rate != impulse_segment.sample_rate: - raise ValueError("Impulse segment's sample rate (%d Hz) is not " - "equal to base signal sample rate (%d Hz)." % - (impulse_segment.sample_rate, self.sample_rate)) - samples = signal.fftconvolve(self.samples, impulse_segment.samples, - "full") - self._samples = samples - - def convolve_and_normalize(self, impulse_segment, allow_resample=False): - """Convolve and normalize the resulting audio segment so that it - has the same average power as the input signal. - - Note that this is an in-place transformation. - - :param impulse_segment: Impulse response segments. - :type impulse_segment: AudioSegment - :param allow_resample: Indicates whether resampling is allowed when - the impulse_segment has a different sample - rate from this signal. - :type allow_resample: bool - """ - target_db = self.rms_db - self.convolve(impulse_segment, allow_resample=allow_resample) - self.normalize(target_db) - - def add_noise(self, - noise, - snr_dB, - allow_downsampling=False, - max_gain_db=300.0, - rng=None): - """Add the given noise segment at a specific signal-to-noise ratio. - If the noise segment is longer than this segment, a random subsegment - of matching length is sampled from it and used instead. - - Note that this is an in-place transformation. - - :param noise: Noise signal to add. - :type noise: AudioSegment - :param snr_dB: Signal-to-Noise Ratio, in decibels. - :type snr_dB: float - :param allow_downsampling: Whether to allow the noise signal to be - downsampled to match the base signal sample - rate. - :type allow_downsampling: bool - :param max_gain_db: Maximum amount of gain to apply to noise signal - before adding it in. This is to prevent attempting - to apply infinite gain to a zero signal. - :type max_gain_db: float - :param rng: Random number generator state. - :type rng: None|random.Random - :raises ValueError: If the sample rate does not match between the two - audio segments when downsampling is not allowed, or - if the duration of noise segments is shorter than - original audio segments. - """ - rng = random.Random() if rng is None else rng - if allow_downsampling and noise.sample_rate > self.sample_rate: - noise = noise.resample(self.sample_rate) - if noise.sample_rate != self.sample_rate: - raise ValueError("Noise sample rate (%d Hz) is not equal to base " - "signal sample rate (%d Hz)." % (noise.sample_rate, - self.sample_rate)) - if noise.duration < self.duration: - raise ValueError("Noise signal (%f sec) must be at least as long as" - " base signal (%f sec)." % - (noise.duration, self.duration)) - noise_gain_db = min(self.rms_db - noise.rms_db - snr_dB, max_gain_db) - noise_new = copy.deepcopy(noise) - noise_new.random_subsegment(self.duration, rng=rng) - noise_new.gain_db(noise_gain_db) - self.superimpose(noise_new) - - @property - def samples(self): - """Return audio samples. - - :return: Audio samples. - :rtype: ndarray - """ - return self._samples.copy() - - @property - def sample_rate(self): - """Return audio sample rate. - - :return: Audio sample rate. - :rtype: int - """ - return self._sample_rate - - @property - def num_samples(self): - """Return number of samples. - - :return: Number of samples. - :rtype: int - """ - return self._samples.shape[0] - - @property - def duration(self): - """Return audio duration. - - :return: Audio duration in seconds. - :rtype: float - """ - return self._samples.shape[0] / float(self._sample_rate) - - @property - def rms_db(self): - """Return root mean square energy of the audio in decibels. - - :return: Root mean square energy in decibels. - :rtype: float - """ - # square root => multiply by 10 instead of 20 for dBs - mean_square = np.mean(self._samples**2) - return 10 * np.log10(mean_square) - - def _convert_samples_to_float32(self, samples): - """Convert sample type to float32. - - Audio sample type is usually integer or float-point. - Integers will be scaled to [-1, 1] in float32. - """ - float32_samples = samples.astype('float32') - if samples.dtype in np.sctypes['int']: - bits = np.iinfo(samples.dtype).bits - float32_samples *= (1. / 2**(bits - 1)) - elif samples.dtype in np.sctypes['float']: - pass - else: - raise TypeError("Unsupported sample type: %s." % samples.dtype) - return float32_samples - - def _convert_samples_from_float32(self, samples, dtype): - """Convert sample type from float32 to dtype. - - Audio sample type is usually integer or float-point. For integer - type, float32 will be rescaled from [-1, 1] to the maximum range - supported by the integer type. - - This is for writing a audio file. - """ - dtype = np.dtype(dtype) - output_samples = samples.copy() - if dtype in np.sctypes['int']: - bits = np.iinfo(dtype).bits - output_samples *= (2**(bits - 1) / 1.) - min_val = np.iinfo(dtype).min - max_val = np.iinfo(dtype).max - output_samples[output_samples > max_val] = max_val - output_samples[output_samples < min_val] = min_val - elif samples.dtype in np.sctypes['float']: - min_val = np.finfo(dtype).min - max_val = np.finfo(dtype).max - output_samples[output_samples > max_val] = max_val - output_samples[output_samples < min_val] = min_val - else: - raise TypeError("Unsupported sample type: %s." % samples.dtype) - return output_samples.astype(dtype) diff --git a/deep_speech_2/data_utils/augmentor/__init__.py b/deep_speech_2/data_utils/augmentor/__init__.py deleted file mode 100644 index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0000000000000000000000000000000000000000 diff --git a/deep_speech_2/data_utils/augmentor/augmentation.py b/deep_speech_2/data_utils/augmentor/augmentation.py deleted file mode 100644 index 5c30b627ef9a23ff41d1f64f270934f149a793a2..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/augmentation.py +++ /dev/null @@ -1,124 +0,0 @@ -"""Contains the data augmentation pipeline.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import json -import random -from data_utils.augmentor.volume_perturb import VolumePerturbAugmentor -from data_utils.augmentor.shift_perturb import ShiftPerturbAugmentor -from data_utils.augmentor.speed_perturb import SpeedPerturbAugmentor -from data_utils.augmentor.noise_perturb import NoisePerturbAugmentor -from data_utils.augmentor.impulse_response import ImpulseResponseAugmentor -from data_utils.augmentor.resample import ResampleAugmentor -from data_utils.augmentor.online_bayesian_normalization import \ - OnlineBayesianNormalizationAugmentor - - -class AugmentationPipeline(object): - """Build a pre-processing pipeline with various augmentation models.Such a - data augmentation pipeline is oftern leveraged to augment the training - samples to make the model invariant to certain types of perturbations in the - real world, improving model's generalization ability. - - The pipeline is built according the the augmentation configuration in json - string, e.g. - - .. code-block:: - - [ { - "type": "noise", - "params": {"min_snr_dB": 10, - "max_snr_dB": 20, - "noise_manifest_path": "datasets/manifest.noise"}, - "prob": 0.0 - }, - { - "type": "speed", - "params": {"min_speed_rate": 0.9, - "max_speed_rate": 1.1}, - "prob": 1.0 - }, - { - "type": "shift", - "params": {"min_shift_ms": -5, - "max_shift_ms": 5}, - "prob": 1.0 - }, - { - "type": "volume", - "params": {"min_gain_dBFS": -10, - "max_gain_dBFS": 10}, - "prob": 0.0 - }, - { - "type": "bayesian_normal", - "params": {"target_db": -20, - "prior_db": -20, - "prior_samples": 100}, - "prob": 0.0 - } - ] - - This augmentation configuration inserts two augmentation models - into the pipeline, with one is VolumePerturbAugmentor and the other - SpeedPerturbAugmentor. "prob" indicates the probability of the current - augmentor to take effect. If "prob" is zero, the augmentor does not take - effect. - - :param augmentation_config: Augmentation configuration in json string. - :type augmentation_config: str - :param random_seed: Random seed. - :type random_seed: int - :raises ValueError: If the augmentation json config is in incorrect format". - """ - - def __init__(self, augmentation_config, random_seed=0): - self._rng = random.Random(random_seed) - self._augmentors, self._rates = self._parse_pipeline_from( - augmentation_config) - - def transform_audio(self, audio_segment): - """Run the pre-processing pipeline for data augmentation. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to process. - :type audio_segment: AudioSegmenet|SpeechSegment - """ - for augmentor, rate in zip(self._augmentors, self._rates): - if self._rng.uniform(0., 1.) < rate: - augmentor.transform_audio(audio_segment) - - def _parse_pipeline_from(self, config_json): - """Parse the config json to build a augmentation pipelien.""" - try: - configs = json.loads(config_json) - augmentors = [ - self._get_augmentor(config["type"], config["params"]) - for config in configs - ] - rates = [config["prob"] for config in configs] - except Exception as e: - raise ValueError("Failed to parse the augmentation config json: " - "%s" % str(e)) - return augmentors, rates - - def _get_augmentor(self, augmentor_type, params): - """Return an augmentation model by the type name, and pass in params.""" - if augmentor_type == "volume": - return VolumePerturbAugmentor(self._rng, **params) - elif augmentor_type == "shift": - return ShiftPerturbAugmentor(self._rng, **params) - elif augmentor_type == "speed": - return SpeedPerturbAugmentor(self._rng, **params) - elif augmentor_type == "resample": - return ResampleAugmentor(self._rng, **params) - elif augmentor_type == "bayesian_normal": - return OnlineBayesianNormalizationAugmentor(self._rng, **params) - elif augmentor_type == "noise": - return NoisePerturbAugmentor(self._rng, **params) - elif augmentor_type == "impulse": - return ImpulseResponseAugmentor(self._rng, **params) - else: - raise ValueError("Unknown augmentor type [%s]." % augmentor_type) diff --git a/deep_speech_2/data_utils/augmentor/base.py b/deep_speech_2/data_utils/augmentor/base.py deleted file mode 100644 index a323165aaeefb8135e7189a47a388a565afd8c8a..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/base.py +++ /dev/null @@ -1,33 +0,0 @@ -"""Contains the abstract base class for augmentation models.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from abc import ABCMeta, abstractmethod - - -class AugmentorBase(object): - """Abstract base class for augmentation model (augmentor) class. - All augmentor classes should inherit from this class, and implement the - following abstract methods. - """ - - __metaclass__ = ABCMeta - - @abstractmethod - def __init__(self): - pass - - @abstractmethod - def transform_audio(self, audio_segment): - """Adds various effects to the input audio segment. Such effects - will augment the training data to make the model invariant to certain - types of perturbations in the real world, improving model's - generalization ability. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to add effects to. - :type audio_segment: AudioSegmenet|SpeechSegment - """ - pass diff --git a/deep_speech_2/data_utils/augmentor/impulse_response.py b/deep_speech_2/data_utils/augmentor/impulse_response.py deleted file mode 100644 index 536b4d6a4a6666359b90e191a3d593250b44e863..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/impulse_response.py +++ /dev/null @@ -1,34 +0,0 @@ -"""Contains the impulse response augmentation model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.augmentor.base import AugmentorBase -from data_utils.utility import read_manifest -from data_utils.audio import AudioSegment - - -class ImpulseResponseAugmentor(AugmentorBase): - """Augmentation model for adding impulse response effect. - - :param rng: Random generator object. - :type rng: random.Random - :param impulse_manifest_path: Manifest path for impulse audio data. - :type impulse_manifest_path: basestring - """ - - def __init__(self, rng, impulse_manifest_path): - self._rng = rng - self._impulse_manifest = read_manifest(impulse_manifest_path) - - def transform_audio(self, audio_segment): - """Add impulse response effect. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to add effects to. - :type audio_segment: AudioSegmenet|SpeechSegment - """ - impulse_json = self._rng.sample(self._impulse_manifest, 1)[0] - impulse_segment = AudioSegment.from_file(impulse_json['audio_filepath']) - audio_segment.convolve(impulse_segment, allow_resample=True) diff --git a/deep_speech_2/data_utils/augmentor/noise_perturb.py b/deep_speech_2/data_utils/augmentor/noise_perturb.py deleted file mode 100644 index 96e0ff4deac48063faf76338014e418e3d8ad4ad..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/noise_perturb.py +++ /dev/null @@ -1,49 +0,0 @@ -"""Contains the noise perturb augmentation model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.augmentor.base import AugmentorBase -from data_utils.utility import read_manifest -from data_utils.audio import AudioSegment - - -class NoisePerturbAugmentor(AugmentorBase): - """Augmentation model for adding background noise. - - :param rng: Random generator object. - :type rng: random.Random - :param min_snr_dB: Minimal signal noise ratio, in decibels. - :type min_snr_dB: float - :param max_snr_dB: Maximal signal noise ratio, in decibels. - :type max_snr_dB: float - :param noise_manifest_path: Manifest path for noise audio data. - :type noise_manifest_path: basestring - """ - - def __init__(self, rng, min_snr_dB, max_snr_dB, noise_manifest_path): - self._min_snr_dB = min_snr_dB - self._max_snr_dB = max_snr_dB - self._rng = rng - self._noise_manifest = read_manifest(manifest_path=noise_manifest_path) - - def transform_audio(self, audio_segment): - """Add background noise audio. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to add effects to. - :type audio_segment: AudioSegmenet|SpeechSegment - """ - noise_json = self._rng.sample(self._noise_manifest, 1)[0] - if noise_json['duration'] < audio_segment.duration: - raise RuntimeError("The duration of sampled noise audio is smaller " - "than the audio segment to add effects to.") - diff_duration = noise_json['duration'] - audio_segment.duration - start = self._rng.uniform(0, diff_duration) - end = start + audio_segment.duration - noise_segment = AudioSegment.slice_from_file( - noise_json['audio_filepath'], start=start, end=end) - snr_dB = self._rng.uniform(self._min_snr_dB, self._max_snr_dB) - audio_segment.add_noise( - noise_segment, snr_dB, allow_downsampling=True, rng=self._rng) diff --git a/deep_speech_2/data_utils/augmentor/online_bayesian_normalization.py b/deep_speech_2/data_utils/augmentor/online_bayesian_normalization.py deleted file mode 100644 index e488ac7d67833631919f88b9e660a99b363b90d0..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/online_bayesian_normalization.py +++ /dev/null @@ -1,48 +0,0 @@ -"""Contain the online bayesian normalization augmentation model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.augmentor.base import AugmentorBase - - -class OnlineBayesianNormalizationAugmentor(AugmentorBase): - """Augmentation model for adding online bayesian normalization. - - :param rng: Random generator object. - :type rng: random.Random - :param target_db: Target RMS value in decibels. - :type target_db: float - :param prior_db: Prior RMS estimate in decibels. - :type prior_db: float - :param prior_samples: Prior strength in number of samples. - :type prior_samples: int - :param startup_delay: Default 0.0s. If provided, this function will - accrue statistics for the first startup_delay - seconds before applying online normalization. - :type starup_delay: float. - """ - - def __init__(self, - rng, - target_db, - prior_db, - prior_samples, - startup_delay=0.0): - self._target_db = target_db - self._prior_db = prior_db - self._prior_samples = prior_samples - self._rng = rng - self._startup_delay = startup_delay - - def transform_audio(self, audio_segment): - """Normalizes the input audio using the online Bayesian approach. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to add effects to. - :type audio_segment: AudioSegment|SpeechSegment - """ - audio_segment.normalize_online_bayesian(self._target_db, self._prior_db, - self._prior_samples, - self._startup_delay) diff --git a/deep_speech_2/data_utils/augmentor/resample.py b/deep_speech_2/data_utils/augmentor/resample.py deleted file mode 100644 index 8df17f3a869420bca1e4e6c0ae9b4035f7d50d8d..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/resample.py +++ /dev/null @@ -1,33 +0,0 @@ -"""Contain the resample augmentation model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.augmentor.base import AugmentorBase - - -class ResampleAugmentor(AugmentorBase): - """Augmentation model for resampling. - - See more info here: - https://ccrma.stanford.edu/~jos/resample/index.html - - :param rng: Random generator object. - :type rng: random.Random - :param new_sample_rate: New sample rate in Hz. - :type new_sample_rate: int - """ - - def __init__(self, rng, new_sample_rate): - self._new_sample_rate = new_sample_rate - self._rng = rng - - def transform_audio(self, audio_segment): - """Resamples the input audio to a target sample rate. - - Note that this is an in-place transformation. - - :param audio: Audio segment to add effects to. - :type audio: AudioSegment|SpeechSegment - """ - audio_segment.resample(self._new_sample_rate) diff --git a/deep_speech_2/data_utils/augmentor/shift_perturb.py b/deep_speech_2/data_utils/augmentor/shift_perturb.py deleted file mode 100644 index c4cbe3e172f6b291f3b778b748affda0341a3181..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/shift_perturb.py +++ /dev/null @@ -1,34 +0,0 @@ -"""Contains the volume perturb augmentation model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.augmentor.base import AugmentorBase - - -class ShiftPerturbAugmentor(AugmentorBase): - """Augmentation model for adding random shift perturbation. - - :param rng: Random generator object. - :type rng: random.Random - :param min_shift_ms: Minimal shift in milliseconds. - :type min_shift_ms: float - :param max_shift_ms: Maximal shift in milliseconds. - :type max_shift_ms: float - """ - - def __init__(self, rng, min_shift_ms, max_shift_ms): - self._min_shift_ms = min_shift_ms - self._max_shift_ms = max_shift_ms - self._rng = rng - - def transform_audio(self, audio_segment): - """Shift audio. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to add effects to. - :type audio_segment: AudioSegmenet|SpeechSegment - """ - shift_ms = self._rng.uniform(self._min_shift_ms, self._max_shift_ms) - audio_segment.shift(shift_ms) diff --git a/deep_speech_2/data_utils/augmentor/speed_perturb.py b/deep_speech_2/data_utils/augmentor/speed_perturb.py deleted file mode 100644 index cc5738bd155a5871817039f5ccb3c4707ff87a6c..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/speed_perturb.py +++ /dev/null @@ -1,47 +0,0 @@ -"""Contain the speech perturbation augmentation model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.augmentor.base import AugmentorBase - - -class SpeedPerturbAugmentor(AugmentorBase): - """Augmentation model for adding speed perturbation. - - See reference paper here: - http://www.danielpovey.com/files/2015_interspeech_augmentation.pdf - - :param rng: Random generator object. - :type rng: random.Random - :param min_speed_rate: Lower bound of new speed rate to sample and should - not be smaller than 0.9. - :type min_speed_rate: float - :param max_speed_rate: Upper bound of new speed rate to sample and should - not be larger than 1.1. - :type max_speed_rate: float - """ - - def __init__(self, rng, min_speed_rate, max_speed_rate): - if min_speed_rate < 0.9: - raise ValueError( - "Sampling speed below 0.9 can cause unnatural effects") - if max_speed_rate > 1.1: - raise ValueError( - "Sampling speed above 1.1 can cause unnatural effects") - self._min_speed_rate = min_speed_rate - self._max_speed_rate = max_speed_rate - self._rng = rng - - def transform_audio(self, audio_segment): - """Sample a new speed rate from the given range and - changes the speed of the given audio clip. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to add effects to. - :type audio_segment: AudioSegment|SpeechSegment - """ - sampled_speed = self._rng.uniform(self._min_speed_rate, - self._max_speed_rate) - audio_segment.change_speed(sampled_speed) diff --git a/deep_speech_2/data_utils/augmentor/volume_perturb.py b/deep_speech_2/data_utils/augmentor/volume_perturb.py deleted file mode 100644 index 758676d558d8e4d77191504d0d7b75cefe020549..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/augmentor/volume_perturb.py +++ /dev/null @@ -1,40 +0,0 @@ -"""Contains the volume perturb augmentation model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.augmentor.base import AugmentorBase - - -class VolumePerturbAugmentor(AugmentorBase): - """Augmentation model for adding random volume perturbation. - - This is used for multi-loudness training of PCEN. See - - https://arxiv.org/pdf/1607.05666v1.pdf - - for more details. - - :param rng: Random generator object. - :type rng: random.Random - :param min_gain_dBFS: Minimal gain in dBFS. - :type min_gain_dBFS: float - :param max_gain_dBFS: Maximal gain in dBFS. - :type max_gain_dBFS: float - """ - - def __init__(self, rng, min_gain_dBFS, max_gain_dBFS): - self._min_gain_dBFS = min_gain_dBFS - self._max_gain_dBFS = max_gain_dBFS - self._rng = rng - - def transform_audio(self, audio_segment): - """Change audio loadness. - - Note that this is an in-place transformation. - - :param audio_segment: Audio segment to add effects to. - :type audio_segment: AudioSegmenet|SpeechSegment - """ - gain = self._rng.uniform(self._min_gain_dBFS, self._max_gain_dBFS) - audio_segment.gain_db(gain) diff --git a/deep_speech_2/data_utils/data.py b/deep_speech_2/data_utils/data.py deleted file mode 100644 index 9dd2a91f639ec08fb2cebc0a91840943ccea6d84..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/data.py +++ /dev/null @@ -1,384 +0,0 @@ -"""Contains data generator for orgnaizing various audio data preprocessing -pipeline and offering data reader interface of PaddlePaddle requirements. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import random -import tarfile -import multiprocessing -import numpy as np -import paddle.v2 as paddle -from threading import local -import atexit -from data_utils.utility import read_manifest -from data_utils.utility import xmap_readers_mp -from data_utils.augmentor.augmentation import AugmentationPipeline -from data_utils.featurizer.speech_featurizer import SpeechFeaturizer -from data_utils.speech import SpeechSegment -from data_utils.normalizer import FeatureNormalizer - - -class DataGenerator(object): - """ - DataGenerator provides basic audio data preprocessing pipeline, and offers - data reader interfaces of PaddlePaddle requirements. - - :param vocab_filepath: Vocabulary filepath for indexing tokenized - transcripts. - :type vocab_filepath: basestring - :param mean_std_filepath: File containing the pre-computed mean and stddev. - :type mean_std_filepath: None|basestring - :param augmentation_config: Augmentation configuration in json string. - Details see AugmentationPipeline.__doc__. - :type augmentation_config: str - :param max_duration: Audio with duration (in seconds) greater than - this will be discarded. - :type max_duration: float - :param min_duration: Audio with duration (in seconds) smaller than - this will be discarded. - :type min_duration: float - :param stride_ms: Striding size (in milliseconds) for generating frames. - :type stride_ms: float - :param window_ms: Window size (in milliseconds) for generating frames. - :type window_ms: float - :param max_freq: Used when specgram_type is 'linear', only FFT bins - corresponding to frequencies between [0, max_freq] are - returned. - :types max_freq: None|float - :param specgram_type: Specgram feature type. Options: 'linear'. - :type specgram_type: str - :param use_dB_normalization: Whether to normalize the audio to -20 dB - before extracting the features. - :type use_dB_normalization: bool - :param num_threads: Number of CPU threads for processing data. - :type num_threads: int - :param random_seed: Random seed. - :type random_seed: int - :param keep_transcription_text: If set to True, transcription text will - be passed forward directly without - converting to index sequence. - :type keep_transcription_text: bool - :param num_conv_layers: The number of convolution layer, used to compute - the sequence length. - :type num_conv_layers: int - """ - - def __init__(self, - vocab_filepath, - mean_std_filepath, - augmentation_config='{}', - max_duration=float('inf'), - min_duration=0.0, - stride_ms=10.0, - window_ms=20.0, - max_freq=None, - specgram_type='linear', - use_dB_normalization=True, - num_threads=multiprocessing.cpu_count() // 2, - random_seed=0, - keep_transcription_text=False, - num_conv_layers=2): - self._max_duration = max_duration - self._min_duration = min_duration - self._normalizer = FeatureNormalizer(mean_std_filepath) - self._augmentation_pipeline = AugmentationPipeline( - augmentation_config=augmentation_config, random_seed=random_seed) - self._speech_featurizer = SpeechFeaturizer( - vocab_filepath=vocab_filepath, - specgram_type=specgram_type, - stride_ms=stride_ms, - window_ms=window_ms, - max_freq=max_freq, - use_dB_normalization=use_dB_normalization) - self._num_threads = num_threads - self._rng = random.Random(random_seed) - self._keep_transcription_text = keep_transcription_text - self._epoch = 0 - # for caching tar files info - self._local_data = local() - self._local_data.tar2info = {} - self._local_data.tar2object = {} - self._num_conv_layers = num_conv_layers - - def process_utterance(self, filename, transcript): - """Load, augment, featurize and normalize for speech data. - - :param filename: Audio filepath - :type filename: basestring | file - :param transcript: Transcription text. - :type transcript: basestring - :return: Tuple of audio feature tensor and data of transcription part, - where transcription part could be token ids or text. - :rtype: tuple of (2darray, list) - """ - if filename.startswith('tar:'): - speech_segment = SpeechSegment.from_file( - self._subfile_from_tar(filename), transcript) - else: - speech_segment = SpeechSegment.from_file(filename, transcript) - self._augmentation_pipeline.transform_audio(speech_segment) - specgram, transcript_part = self._speech_featurizer.featurize( - speech_segment, self._keep_transcription_text) - specgram = self._normalizer.apply(specgram) - return specgram, transcript_part - - def batch_reader_creator(self, - manifest_path, - batch_size, - min_batch_size=1, - padding_to=-1, - flatten=False, - sortagrad=False, - shuffle_method="batch_shuffle"): - """ - Batch data reader creator for audio data. Return a callable generator - function to produce batches of data. - - Audio features within one batch will be padded with zeros to have the - same shape, or a user-defined shape. - - :param manifest_path: Filepath of manifest for audio files. - :type manifest_path: basestring - :param batch_size: Number of instances in a batch. - :type batch_size: int - :param min_batch_size: Any batch with batch size smaller than this will - be discarded. (To be deprecated in the future.) - :type min_batch_size: int - :param padding_to: If set -1, the maximun shape in the batch - will be used as the target shape for padding. - Otherwise, `padding_to` will be the target shape. - :type padding_to: int - :param flatten: If set True, audio features will be flatten to 1darray. - :type flatten: bool - :param sortagrad: If set True, sort the instances by audio duration - in the first epoch for speed up training. - :type sortagrad: bool - :param shuffle_method: Shuffle method. Options: - '' or None: no shuffle. - 'instance_shuffle': instance-wise shuffle. - 'batch_shuffle': similarly-sized instances are - put into batches, and then - batch-wise shuffle the batches. - For more details, please see - ``_batch_shuffle.__doc__``. - 'batch_shuffle_clipped': 'batch_shuffle' with - head shift and tail - clipping. For more - details, please see - ``_batch_shuffle``. - If sortagrad is True, shuffle is disabled - for the first epoch. - :type shuffle_method: None|str - :return: Batch reader function, producing batches of data when called. - :rtype: callable - """ - - def batch_reader(): - # read manifest - manifest = read_manifest( - manifest_path=manifest_path, - max_duration=self._max_duration, - min_duration=self._min_duration) - # sort (by duration) or batch-wise shuffle the manifest - if self._epoch == 0 and sortagrad: - manifest.sort(key=lambda x: x["duration"]) - else: - if shuffle_method == "batch_shuffle": - manifest = self._batch_shuffle( - manifest, batch_size, clipped=False) - elif shuffle_method == "batch_shuffle_clipped": - manifest = self._batch_shuffle( - manifest, batch_size, clipped=True) - elif shuffle_method == "instance_shuffle": - self._rng.shuffle(manifest) - elif shuffle_method == None: - pass - else: - raise ValueError("Unknown shuffle method %s." % - shuffle_method) - # prepare batches - instance_reader = self._instance_reader_creator(manifest) - batch = [] - for instance in instance_reader(): - batch.append(instance) - if len(batch) == batch_size: - yield self._padding_batch(batch, padding_to, flatten) - batch = [] - if len(batch) >= min_batch_size: - yield self._padding_batch(batch, padding_to, flatten) - self._epoch += 1 - - return batch_reader - - @property - def feeding(self): - """Returns data reader's feeding dict. - - :return: Data feeding dict. - :rtype: dict - """ - feeding_dict = { - "audio_spectrogram": 0, - "transcript_text": 1, - "sequence_offset": 2, - "sequence_length": 3 - } - for i in xrange(self._num_conv_layers): - feeding_dict["conv%d_index_range" % i] = len(feeding_dict) - return feeding_dict - - @property - def vocab_size(self): - """Return the vocabulary size. - - :return: Vocabulary size. - :rtype: int - """ - return self._speech_featurizer.vocab_size - - @property - def vocab_list(self): - """Return the vocabulary in list. - - :return: Vocabulary in list. - :rtype: list - """ - return self._speech_featurizer.vocab_list - - def _parse_tar(self, file): - """Parse a tar file to get a tarfile object - and a map containing tarinfoes - """ - result = {} - f = tarfile.open(file) - for tarinfo in f.getmembers(): - result[tarinfo.name] = tarinfo - return f, result - - def _subfile_from_tar(self, file): - """Get subfile object from tar. - - It will return a subfile object from tar file - and cached tar file info for next reading request. - """ - tarpath, filename = file.split(':', 1)[1].split('#', 1) - if 'tar2info' not in self._local_data.__dict__: - self._local_data.tar2info = {} - if 'tar2object' not in self._local_data.__dict__: - self._local_data.tar2object = {} - if tarpath not in self._local_data.tar2info: - object, infoes = self._parse_tar(tarpath) - self._local_data.tar2info[tarpath] = infoes - self._local_data.tar2object[tarpath] = object - return self._local_data.tar2object[tarpath].extractfile( - self._local_data.tar2info[tarpath][filename]) - - def _instance_reader_creator(self, manifest): - """ - Instance reader creator. Create a callable function to produce - instances of data. - - Instance: a tuple of ndarray of audio spectrogram and a list of - token indices for transcript. - """ - - def reader(): - for instance in manifest: - yield instance - - reader, cleanup_callback = xmap_readers_mp( - lambda instance: self.process_utterance(instance["audio_filepath"], instance["text"]), - reader, - self._num_threads, - 4096, - order=True) - - # register callback to main process - atexit.register(cleanup_callback) - - return reader - - def _padding_batch(self, batch, padding_to=-1, flatten=False): - """ - Padding audio features with zeros to make them have the same shape (or - a user-defined shape) within one bach. - - If ``padding_to`` is -1, the maximun shape in the batch will be used - as the target shape for padding. Otherwise, `padding_to` will be the - target shape (only refers to the second axis). - - If `flatten` is True, features will be flatten to 1darray. - """ - new_batch = [] - # get target shape - max_length = max([audio.shape[1] for audio, text in batch]) - if padding_to != -1: - if padding_to < max_length: - raise ValueError("If padding_to is not -1, it should be larger " - "than any instance's shape in the batch") - max_length = padding_to - # padding - for audio, text in batch: - padded_audio = np.zeros([audio.shape[0], max_length]) - padded_audio[:, :audio.shape[1]] = audio - if flatten: - padded_audio = padded_audio.flatten() - - # Stride size for conv0 is (3, 2) - # Stride size for conv1 to convN is (1, 2) - # Same as the network, hard-coded here - padded_instance = [padded_audio, text] - padded_conv0_h = (padded_audio.shape[0] - 1) // 2 + 1 - padded_conv0_w = (padded_audio.shape[1] - 1) // 3 + 1 - valid_w = (audio.shape[1] - 1) // 3 + 1 - padded_instance += [ - [0], # sequence offset, always 0 - [valid_w], # valid sequence length - # Index ranges for channel, height and width - # Please refer scale_sub_region layer to see details - [1, 32, 1, padded_conv0_h, valid_w + 1, padded_conv0_w] - ] - pre_padded_h = padded_conv0_h - for i in xrange(self._num_conv_layers - 1): - padded_h = (pre_padded_h - 1) // 2 + 1 - pre_padded_h = padded_h - padded_instance += [ - [1, 32, 1, padded_h, valid_w + 1, padded_conv0_w] - ] - - new_batch.append(padded_instance) - return new_batch - - def _batch_shuffle(self, manifest, batch_size, clipped=False): - """Put similarly-sized instances into minibatches for better efficiency - and make a batch-wise shuffle. - - 1. Sort the audio clips by duration. - 2. Generate a random number `k`, k in [0, batch_size). - 3. Randomly shift `k` instances in order to create different batches - for different epochs. Create minibatches. - 4. Shuffle the minibatches. - - :param manifest: Manifest contents. List of dict. - :type manifest: list - :param batch_size: Batch size. This size is also used for generate - a random number for batch shuffle. - :type batch_size: int - :param clipped: Whether to clip the heading (small shift) and trailing - (incomplete batch) instances. - :type clipped: bool - :return: Batch shuffled mainifest. - :rtype: list - """ - manifest.sort(key=lambda x: x["duration"]) - shift_len = self._rng.randint(0, batch_size - 1) - batch_manifest = zip(*[iter(manifest[shift_len:])] * batch_size) - self._rng.shuffle(batch_manifest) - batch_manifest = [item for batch in batch_manifest for item in batch] - if not clipped: - res_len = len(manifest) - shift_len - len(batch_manifest) - batch_manifest.extend(manifest[-res_len:]) - batch_manifest.extend(manifest[0:shift_len]) - return batch_manifest diff --git a/deep_speech_2/data_utils/featurizer/__init__.py b/deep_speech_2/data_utils/featurizer/__init__.py deleted file mode 100644 index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0000000000000000000000000000000000000000 diff --git a/deep_speech_2/data_utils/featurizer/audio_featurizer.py b/deep_speech_2/data_utils/featurizer/audio_featurizer.py deleted file mode 100644 index f594de7d9794354ed4a5afca25d70e85d658a828..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/featurizer/audio_featurizer.py +++ /dev/null @@ -1,187 +0,0 @@ -"""Contains the audio featurizer class.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import numpy as np -from data_utils.utility import read_manifest -from data_utils.audio import AudioSegment -from python_speech_features import mfcc -from python_speech_features import delta - - -class AudioFeaturizer(object): - """Audio featurizer, for extracting features from audio contents of - AudioSegment or SpeechSegment. - - Currently, it supports feature types of linear spectrogram and mfcc. - - :param specgram_type: Specgram feature type. Options: 'linear'. - :type specgram_type: str - :param stride_ms: Striding size (in milliseconds) for generating frames. - :type stride_ms: float - :param window_ms: Window size (in milliseconds) for generating frames. - :type window_ms: float - :param max_freq: When specgram_type is 'linear', only FFT bins - corresponding to frequencies between [0, max_freq] are - returned; when specgram_type is 'mfcc', max_feq is the - highest band edge of mel filters. - :types max_freq: None|float - :param target_sample_rate: Audio are resampled (if upsampling or - downsampling is allowed) to this before - extracting spectrogram features. - :type target_sample_rate: float - :param use_dB_normalization: Whether to normalize the audio to a certain - decibels before extracting the features. - :type use_dB_normalization: bool - :param target_dB: Target audio decibels for normalization. - :type target_dB: float - """ - - def __init__(self, - specgram_type='linear', - stride_ms=10.0, - window_ms=20.0, - max_freq=None, - target_sample_rate=16000, - use_dB_normalization=True, - target_dB=-20): - self._specgram_type = specgram_type - self._stride_ms = stride_ms - self._window_ms = window_ms - self._max_freq = max_freq - self._target_sample_rate = target_sample_rate - self._use_dB_normalization = use_dB_normalization - self._target_dB = target_dB - - def featurize(self, - audio_segment, - allow_downsampling=True, - allow_upsampling=True): - """Extract audio features from AudioSegment or SpeechSegment. - - :param audio_segment: Audio/speech segment to extract features from. - :type audio_segment: AudioSegment|SpeechSegment - :param allow_downsampling: Whether to allow audio downsampling before - featurizing. - :type allow_downsampling: bool - :param allow_upsampling: Whether to allow audio upsampling before - featurizing. - :type allow_upsampling: bool - :return: Spectrogram audio feature in 2darray. - :rtype: ndarray - :raises ValueError: If audio sample rate is not supported. - """ - # upsampling or downsampling - if ((audio_segment.sample_rate > self._target_sample_rate and - allow_downsampling) or - (audio_segment.sample_rate < self._target_sample_rate and - allow_upsampling)): - audio_segment.resample(self._target_sample_rate) - if audio_segment.sample_rate != self._target_sample_rate: - raise ValueError("Audio sample rate is not supported. " - "Turn allow_downsampling or allow up_sampling on.") - # decibel normalization - if self._use_dB_normalization: - audio_segment.normalize(target_db=self._target_dB) - # extract spectrogram - return self._compute_specgram(audio_segment.samples, - audio_segment.sample_rate) - - def _compute_specgram(self, samples, sample_rate): - """Extract various audio features.""" - if self._specgram_type == 'linear': - return self._compute_linear_specgram( - samples, sample_rate, self._stride_ms, self._window_ms, - self._max_freq) - elif self._specgram_type == 'mfcc': - return self._compute_mfcc(samples, sample_rate, self._stride_ms, - self._window_ms, self._max_freq) - else: - raise ValueError("Unknown specgram_type %s. " - "Supported values: linear." % self._specgram_type) - - def _compute_linear_specgram(self, - samples, - sample_rate, - stride_ms=10.0, - window_ms=20.0, - max_freq=None, - eps=1e-14): - """Compute the linear spectrogram from FFT energy.""" - if max_freq is None: - max_freq = sample_rate / 2 - if max_freq > sample_rate / 2: - raise ValueError("max_freq must be greater than half of " - "sample rate.") - if stride_ms > window_ms: - raise ValueError("Stride size must not be greater than " - "window size.") - stride_size = int(0.001 * sample_rate * stride_ms) - window_size = int(0.001 * sample_rate * window_ms) - specgram, freqs = self._specgram_real( - samples, - window_size=window_size, - stride_size=stride_size, - sample_rate=sample_rate) - ind = np.where(freqs <= max_freq)[0][-1] + 1 - return np.log(specgram[:ind, :] + eps) - - def _specgram_real(self, samples, window_size, stride_size, sample_rate): - """Compute the spectrogram for samples from a real signal.""" - # extract strided windows - truncate_size = (len(samples) - window_size) % stride_size - samples = samples[:len(samples) - truncate_size] - nshape = (window_size, (len(samples) - window_size) // stride_size + 1) - nstrides = (samples.strides[0], samples.strides[0] * stride_size) - windows = np.lib.stride_tricks.as_strided( - samples, shape=nshape, strides=nstrides) - assert np.all( - windows[:, 1] == samples[stride_size:(stride_size + window_size)]) - # window weighting, squared Fast Fourier Transform (fft), scaling - weighting = np.hanning(window_size)[:, None] - fft = np.fft.rfft(windows * weighting, axis=0) - fft = np.absolute(fft) - fft = fft**2 - scale = np.sum(weighting**2) * sample_rate - fft[1:-1, :] *= (2.0 / scale) - fft[(0, -1), :] /= scale - # prepare fft frequency list - freqs = float(sample_rate) / window_size * np.arange(fft.shape[0]) - return fft, freqs - - def _compute_mfcc(self, - samples, - sample_rate, - stride_ms=10.0, - window_ms=20.0, - max_freq=None): - """Compute mfcc from samples.""" - if max_freq is None: - max_freq = sample_rate / 2 - if max_freq > sample_rate / 2: - raise ValueError("max_freq must not be greater than half of " - "sample rate.") - if stride_ms > window_ms: - raise ValueError("Stride size must not be greater than " - "window size.") - # compute the 13 cepstral coefficients, and the first one is replaced - # by log(frame energy) - mfcc_feat = mfcc( - signal=samples, - samplerate=sample_rate, - winlen=0.001 * window_ms, - winstep=0.001 * stride_ms, - highfreq=max_freq) - # Deltas - d_mfcc_feat = delta(mfcc_feat, 2) - # Deltas-Deltas - dd_mfcc_feat = delta(d_mfcc_feat, 2) - # transpose - mfcc_feat = np.transpose(mfcc_feat) - d_mfcc_feat = np.transpose(d_mfcc_feat) - dd_mfcc_feat = np.transpose(dd_mfcc_feat) - # concat above three features - concat_mfcc_feat = np.concatenate( - (mfcc_feat, d_mfcc_feat, dd_mfcc_feat)) - return concat_mfcc_feat diff --git a/deep_speech_2/data_utils/featurizer/speech_featurizer.py b/deep_speech_2/data_utils/featurizer/speech_featurizer.py deleted file mode 100644 index 4555dc31da89367b4775b712d3876168aae268f4..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/featurizer/speech_featurizer.py +++ /dev/null @@ -1,98 +0,0 @@ -"""Contains the speech featurizer class.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.featurizer.audio_featurizer import AudioFeaturizer -from data_utils.featurizer.text_featurizer import TextFeaturizer - - -class SpeechFeaturizer(object): - """Speech featurizer, for extracting features from both audio and transcript - contents of SpeechSegment. - - Currently, for audio parts, it supports feature types of linear - spectrogram and mfcc; for transcript parts, it only supports char-level - tokenizing and conversion into a list of token indices. Note that the - token indexing order follows the given vocabulary file. - - :param vocab_filepath: Filepath to load vocabulary for token indices - conversion. - :type specgram_type: basestring - :param specgram_type: Specgram feature type. Options: 'linear', 'mfcc'. - :type specgram_type: str - :param stride_ms: Striding size (in milliseconds) for generating frames. - :type stride_ms: float - :param window_ms: Window size (in milliseconds) for generating frames. - :type window_ms: float - :param max_freq: When specgram_type is 'linear', only FFT bins - corresponding to frequencies between [0, max_freq] are - returned; when specgram_type is 'mfcc', max_freq is the - highest band edge of mel filters. - :types max_freq: None|float - :param target_sample_rate: Speech are resampled (if upsampling or - downsampling is allowed) to this before - extracting spectrogram features. - :type target_sample_rate: float - :param use_dB_normalization: Whether to normalize the audio to a certain - decibels before extracting the features. - :type use_dB_normalization: bool - :param target_dB: Target audio decibels for normalization. - :type target_dB: float - """ - - def __init__(self, - vocab_filepath, - specgram_type='linear', - stride_ms=10.0, - window_ms=20.0, - max_freq=None, - target_sample_rate=16000, - use_dB_normalization=True, - target_dB=-20): - self._audio_featurizer = AudioFeaturizer( - specgram_type=specgram_type, - stride_ms=stride_ms, - window_ms=window_ms, - max_freq=max_freq, - target_sample_rate=target_sample_rate, - use_dB_normalization=use_dB_normalization, - target_dB=target_dB) - self._text_featurizer = TextFeaturizer(vocab_filepath) - - def featurize(self, speech_segment, keep_transcription_text): - """Extract features for speech segment. - - 1. For audio parts, extract the audio features. - 2. For transcript parts, keep the original text or convert text string - to a list of token indices in char-level. - - :param audio_segment: Speech segment to extract features from. - :type audio_segment: SpeechSegment - :return: A tuple of 1) spectrogram audio feature in 2darray, 2) list of - char-level token indices. - :rtype: tuple - """ - audio_feature = self._audio_featurizer.featurize(speech_segment) - if keep_transcription_text: - return audio_feature, speech_segment.transcript - text_ids = self._text_featurizer.featurize(speech_segment.transcript) - return audio_feature, text_ids - - @property - def vocab_size(self): - """Return the vocabulary size. - - :return: Vocabulary size. - :rtype: int - """ - return self._text_featurizer.vocab_size - - @property - def vocab_list(self): - """Return the vocabulary in list. - - :return: Vocabulary in list. - :rtype: list - """ - return self._text_featurizer.vocab_list diff --git a/deep_speech_2/data_utils/featurizer/text_featurizer.py b/deep_speech_2/data_utils/featurizer/text_featurizer.py deleted file mode 100644 index 89202163ca8d8b69f59b858db5451882d7e089b3..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/featurizer/text_featurizer.py +++ /dev/null @@ -1,68 +0,0 @@ -"""Contains the text featurizer class.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os -import codecs - - -class TextFeaturizer(object): - """Text featurizer, for processing or extracting features from text. - - Currently, it only supports char-level tokenizing and conversion into - a list of token indices. Note that the token indexing order follows the - given vocabulary file. - - :param vocab_filepath: Filepath to load vocabulary for token indices - conversion. - :type specgram_type: basestring - """ - - def __init__(self, vocab_filepath): - self._vocab_dict, self._vocab_list = self._load_vocabulary_from_file( - vocab_filepath) - - def featurize(self, text): - """Convert text string to a list of token indices in char-level.Note - that the token indexing order follows the given vocabulary file. - - :param text: Text to process. - :type text: basestring - :return: List of char-level token indices. - :rtype: list - """ - tokens = self._char_tokenize(text) - return [self._vocab_dict[token] for token in tokens] - - @property - def vocab_size(self): - """Return the vocabulary size. - - :return: Vocabulary size. - :rtype: int - """ - return len(self._vocab_list) - - @property - def vocab_list(self): - """Return the vocabulary in list. - - :return: Vocabulary in list. - :rtype: list - """ - return self._vocab_list - - def _char_tokenize(self, text): - """Character tokenizer.""" - return list(text.strip()) - - def _load_vocabulary_from_file(self, vocab_filepath): - """Load vocabulary from file.""" - vocab_lines = [] - with codecs.open(vocab_filepath, 'r', 'utf-8') as file: - vocab_lines.extend(file.readlines()) - vocab_list = [line[:-1] for line in vocab_lines] - vocab_dict = dict( - [(token, id) for (id, token) in enumerate(vocab_list)]) - return vocab_dict, vocab_list diff --git a/deep_speech_2/data_utils/normalizer.py b/deep_speech_2/data_utils/normalizer.py deleted file mode 100644 index 7c2e05c9d85fa55c0a91386ebf9ba570b2ec0e3b..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/normalizer.py +++ /dev/null @@ -1,87 +0,0 @@ -"""Contains feature normalizers.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import numpy as np -import random -from data_utils.utility import read_manifest -from data_utils.audio import AudioSegment - - -class FeatureNormalizer(object): - """Feature normalizer. Normalize features to be of zero mean and unit - stddev. - - if mean_std_filepath is provided (not None), the normalizer will directly - initilize from the file. Otherwise, both manifest_path and featurize_func - should be given for on-the-fly mean and stddev computing. - - :param mean_std_filepath: File containing the pre-computed mean and stddev. - :type mean_std_filepath: None|basestring - :param manifest_path: Manifest of instances for computing mean and stddev. - :type meanifest_path: None|basestring - :param featurize_func: Function to extract features. It should be callable - with ``featurize_func(audio_segment)``. - :type featurize_func: None|callable - :param num_samples: Number of random samples for computing mean and stddev. - :type num_samples: int - :param random_seed: Random seed for sampling instances. - :type random_seed: int - :raises ValueError: If both mean_std_filepath and manifest_path - (or both mean_std_filepath and featurize_func) are None. - """ - - def __init__(self, - mean_std_filepath, - manifest_path=None, - featurize_func=None, - num_samples=500, - random_seed=0): - if not mean_std_filepath: - if not (manifest_path and featurize_func): - raise ValueError("If mean_std_filepath is None, meanifest_path " - "and featurize_func should not be None.") - self._rng = random.Random(random_seed) - self._compute_mean_std(manifest_path, featurize_func, num_samples) - else: - self._read_mean_std_from_file(mean_std_filepath) - - def apply(self, features, eps=1e-14): - """Normalize features to be of zero mean and unit stddev. - - :param features: Input features to be normalized. - :type features: ndarray - :param eps: added to stddev to provide numerical stablibity. - :type eps: float - :return: Normalized features. - :rtype: ndarray - """ - return (features - self._mean) / (self._std + eps) - - def write_to_file(self, filepath): - """Write the mean and stddev to the file. - - :param filepath: File to write mean and stddev. - :type filepath: basestring - """ - np.savez(filepath, mean=self._mean, std=self._std) - - def _read_mean_std_from_file(self, filepath): - """Load mean and std from file.""" - npzfile = np.load(filepath) - self._mean = npzfile["mean"] - self._std = npzfile["std"] - - def _compute_mean_std(self, manifest_path, featurize_func, num_samples): - """Compute mean and std from randomly sampled instances.""" - manifest = read_manifest(manifest_path) - sampled_manifest = self._rng.sample(manifest, num_samples) - features = [] - for instance in sampled_manifest: - features.append( - featurize_func( - AudioSegment.from_file(instance["audio_filepath"]))) - features = np.hstack(features) - self._mean = np.mean(features, axis=1).reshape([-1, 1]) - self._std = np.std(features, axis=1).reshape([-1, 1]) diff --git a/deep_speech_2/data_utils/speech.py b/deep_speech_2/data_utils/speech.py deleted file mode 100644 index 0cea887309d7c443a5dcdb3577ad897e6b36e209..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/speech.py +++ /dev/null @@ -1,143 +0,0 @@ -"""Contains the speech segment class.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from data_utils.audio import AudioSegment - - -class SpeechSegment(AudioSegment): - """Speech segment abstraction, a subclass of AudioSegment, - with an additional transcript. - - :param samples: Audio samples [num_samples x num_channels]. - :type samples: ndarray.float32 - :param sample_rate: Audio sample rate. - :type sample_rate: int - :param transcript: Transcript text for the speech. - :type transript: basestring - :raises TypeError: If the sample data type is not float or int. - """ - - def __init__(self, samples, sample_rate, transcript): - AudioSegment.__init__(self, samples, sample_rate) - self._transcript = transcript - - def __eq__(self, other): - """Return whether two objects are equal. - """ - if not AudioSegment.__eq__(self, other): - return False - if self._transcript != other._transcript: - return False - return True - - def __ne__(self, other): - """Return whether two objects are unequal.""" - return not self.__eq__(other) - - @classmethod - def from_file(cls, filepath, transcript): - """Create speech segment from audio file and corresponding transcript. - - :param filepath: Filepath or file object to audio file. - :type filepath: basestring|file - :param transcript: Transcript text for the speech. - :type transript: basestring - :return: Speech segment instance. - :rtype: SpeechSegment - """ - audio = AudioSegment.from_file(filepath) - return cls(audio.samples, audio.sample_rate, transcript) - - @classmethod - def from_bytes(cls, bytes, transcript): - """Create speech segment from a byte string and corresponding - transcript. - - :param bytes: Byte string containing audio samples. - :type bytes: str - :param transcript: Transcript text for the speech. - :type transript: basestring - :return: Speech segment instance. - :rtype: Speech Segment - """ - audio = AudioSegment.from_bytes(bytes) - return cls(audio.samples, audio.sample_rate, transcript) - - @classmethod - def concatenate(cls, *segments): - """Concatenate an arbitrary number of speech segments together, both - audio and transcript will be concatenated. - - :param *segments: Input speech segments to be concatenated. - :type *segments: tuple of SpeechSegment - :return: Speech segment instance. - :rtype: SpeechSegment - :raises ValueError: If the number of segments is zero, or if the - sample_rate of any two segments does not match. - :raises TypeError: If any segment is not SpeechSegment instance. - """ - if len(segments) == 0: - raise ValueError("No speech segments are given to concatenate.") - sample_rate = segments[0]._sample_rate - transcripts = "" - for seg in segments: - if sample_rate != seg._sample_rate: - raise ValueError("Can't concatenate segments with " - "different sample rates") - if type(seg) is not cls: - raise TypeError("Only speech segments of the same type " - "instance can be concatenated.") - transcripts += seg._transcript - samples = np.concatenate([seg.samples for seg in segments]) - return cls(samples, sample_rate, transcripts) - - @classmethod - def slice_from_file(cls, filepath, transcript, start=None, end=None): - """Loads a small section of an speech without having to load - the entire file into the memory which can be incredibly wasteful. - - :param filepath: Filepath or file object to audio file. - :type filepath: basestring|file - :param start: Start time in seconds. If start is negative, it wraps - around from the end. If not provided, this function - reads from the very beginning. - :type start: float - :param end: End time in seconds. If end is negative, it wraps around - from the end. If not provided, the default behvaior is - to read to the end of the file. - :type end: float - :param transcript: Transcript text for the speech. if not provided, - the defaults is an empty string. - :type transript: basestring - :return: SpeechSegment instance of the specified slice of the input - speech file. - :rtype: SpeechSegment - """ - audio = AudioSegment.slice_from_file(filepath, start, end) - return cls(audio.samples, audio.sample_rate, transcript) - - @classmethod - def make_silence(cls, duration, sample_rate): - """Creates a silent speech segment of the given duration and - sample rate, transcript will be an empty string. - - :param duration: Length of silence in seconds. - :type duration: float - :param sample_rate: Sample rate. - :type sample_rate: float - :return: Silence of the given duration. - :rtype: SpeechSegment - """ - audio = AudioSegment.make_silence(duration, sample_rate) - return cls(audio.samples, audio.sample_rate, "") - - @property - def transcript(self): - """Return the transcript text. - - :return: Transcript text for the speech. - :rtype: basestring - """ - return self._transcript diff --git a/deep_speech_2/data_utils/utility.py b/deep_speech_2/data_utils/utility.py deleted file mode 100644 index bb5cad45bb2383e2125ec22181c9aedea8508eb0..0000000000000000000000000000000000000000 --- a/deep_speech_2/data_utils/utility.py +++ /dev/null @@ -1,181 +0,0 @@ -"""Contains data helper functions.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import json -import codecs -import os -import tarfile -import time -from Queue import Queue -from threading import Thread -from multiprocessing import Process, Manager -from paddle.v2.dataset.common import md5file - - -def read_manifest(manifest_path, max_duration=float('inf'), min_duration=0.0): - """Load and parse manifest file. - - Instances with durations outside [min_duration, max_duration] will be - filtered out. - - :param manifest_path: Manifest file to load and parse. - :type manifest_path: basestring - :param max_duration: Maximal duration in seconds for instance filter. - :type max_duration: float - :param min_duration: Minimal duration in seconds for instance filter. - :type min_duration: float - :return: Manifest parsing results. List of dict. - :rtype: list - :raises IOError: If failed to parse the manifest. - """ - manifest = [] - for json_line in codecs.open(manifest_path, 'r', 'utf-8'): - try: - json_data = json.loads(json_line) - except Exception as e: - raise IOError("Error reading manifest: %s" % str(e)) - if (json_data["duration"] <= max_duration and - json_data["duration"] >= min_duration): - manifest.append(json_data) - return manifest - - -def download(url, md5sum, target_dir): - """Download file from url to target_dir, and check md5sum.""" - if not os.path.exists(target_dir): os.makedirs(target_dir) - filepath = os.path.join(target_dir, url.split("/")[-1]) - if not (os.path.exists(filepath) and md5file(filepath) == md5sum): - print("Downloading %s ..." % url) - os.system("wget -c " + url + " -P " + target_dir) - print("\nMD5 Chesksum %s ..." % filepath) - if not md5file(filepath) == md5sum: - raise RuntimeError("MD5 checksum failed.") - else: - print("File exists, skip downloading. (%s)" % filepath) - return filepath - - -def unpack(filepath, target_dir, rm_tar=False): - """Unpack the file to the target_dir.""" - print("Unpacking %s ..." % filepath) - tar = tarfile.open(filepath) - tar.extractall(target_dir) - tar.close() - if rm_tar == True: - os.remove(filepath) - - -class XmapEndSignal(): - pass - - -def xmap_readers_mp(mapper, reader, process_num, buffer_size, order=False): - """A multiprocessing pipeline wrapper for the data reader. - - :param mapper: Function to map sample. - :type mapper: callable - :param reader: Given data reader. - :type reader: callable - :param process_num: Number of processes in the pipeline - :type process_num: int - :param buffer_size: Maximal buffer size. - :type buffer_size: int - :param order: Reserve the order of samples from the given reader. - :type order: bool - :return: The wrappered reader - :rtype: callable - """ - end_flag = XmapEndSignal() - - # define a worker to read samples from reader to in_queue - def read_worker(reader, in_queue): - for sample in reader(): - in_queue.put(sample) - in_queue.put(end_flag) - - # define a worker to read samples from reader to in_queue with order flag - def order_read_worker(reader, in_queue): - for order_id, sample in enumerate(reader()): - in_queue.put((order_id, sample)) - in_queue.put(end_flag) - - # define a worker to handle samples from in_queue by mapper and put results - # to out_queue - def handle_worker(in_queue, out_queue, mapper): - sample = in_queue.get() - while not isinstance(sample, XmapEndSignal): - out_queue.put(mapper(sample)) - sample = in_queue.get() - in_queue.put(end_flag) - out_queue.put(end_flag) - - # define a worker to handle samples from in_queue by mapper and put results - # to out_queue with order - def order_handle_worker(in_queue, out_queue, mapper, out_order): - ins = in_queue.get() - while not isinstance(ins, XmapEndSignal): - order_id, sample = ins - result = mapper(sample) - while order_id != out_order[0]: - time.sleep(0.001) - out_queue.put(result) - out_order[0] += 1 - ins = in_queue.get() - in_queue.put(end_flag) - out_queue.put(end_flag) - - # define a thread worker to flush samples from Manager.Queue to Queue - # for acceleration - def flush_worker(in_queue, out_queue): - finish = 0 - while finish < process_num: - sample = in_queue.get() - if isinstance(sample, XmapEndSignal): - finish += 1 - else: - out_queue.put(sample) - out_queue.put(end_flag) - - def cleanup(): - # kill all sub process and threads - os._exit(0) - - def xreader(): - # prepare shared memory - manager = Manager() - in_queue = manager.Queue(buffer_size) - out_queue = manager.Queue(buffer_size) - out_order = manager.list([0]) - - # start a read worker in a process - target = order_read_worker if order else read_worker - p = Process(target=target, args=(reader, in_queue)) - p.daemon = True - p.start() - - # start handle_workers with multiple processes - target = order_handle_worker if order else handle_worker - args = (in_queue, out_queue, mapper, out_order) if order else ( - in_queue, out_queue, mapper) - workers = [ - Process(target=target, args=args) for _ in xrange(process_num) - ] - for w in workers: - w.daemon = True - w.start() - - # start a thread to read data from slow Manager.Queue - flush_queue = Queue(buffer_size) - t = Thread(target=flush_worker, args=(out_queue, flush_queue)) - t.daemon = True - t.start() - - # get results - sample = flush_queue.get() - while not isinstance(sample, XmapEndSignal): - yield sample - sample = flush_queue.get() - - return xreader, cleanup diff --git a/deep_speech_2/decoders/__init__.py b/deep_speech_2/decoders/__init__.py deleted file mode 100644 index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0000000000000000000000000000000000000000 diff --git a/deep_speech_2/decoders/decoders_deprecated.py b/deep_speech_2/decoders/decoders_deprecated.py deleted file mode 100644 index 17b28b0d02a22a2e59856156ccd663324e886aed..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/decoders_deprecated.py +++ /dev/null @@ -1,238 +0,0 @@ -"""Contains various CTC decoders.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from itertools import groupby -import numpy as np -from math import log -import multiprocessing - - -def ctc_greedy_decoder(probs_seq, vocabulary): - """CTC greedy (best path) decoder. - - Path consisting of the most probable tokens are further post-processed to - remove consecutive repetitions and all blanks. - - :param probs_seq: 2-D list of probabilities over the vocabulary for each - character. Each element is a list of float probabilities - for one character. - :type probs_seq: list - :param vocabulary: Vocabulary list. - :type vocabulary: list - :return: Decoding result string. - :rtype: baseline - """ - # dimension verification - for probs in probs_seq: - if not len(probs) == len(vocabulary) + 1: - raise ValueError("probs_seq dimension mismatchedd with vocabulary") - # argmax to get the best index for each time step - max_index_list = list(np.array(probs_seq).argmax(axis=1)) - # remove consecutive duplicate indexes - index_list = [index_group[0] for index_group in groupby(max_index_list)] - # remove blank indexes - blank_index = len(vocabulary) - index_list = [index for index in index_list if index != blank_index] - # convert index list to string - return ''.join([vocabulary[index] for index in index_list]) - - -def ctc_beam_search_decoder(probs_seq, - beam_size, - vocabulary, - cutoff_prob=1.0, - cutoff_top_n=40, - ext_scoring_func=None, - nproc=False): - """CTC Beam search decoder. - - It utilizes beam search to approximately select top best decoding - labels and returning results in the descending order. - The implementation is based on Prefix Beam Search - (https://arxiv.org/abs/1408.2873), and the unclear part is - redesigned. Two important modifications: 1) in the iterative computation - of probabilities, the assignment operation is changed to accumulation for - one prefix may comes from different paths; 2) the if condition "if l^+ not - in A_prev then" after probabilities' computation is deprecated for it is - hard to understand and seems unnecessary. - - :param probs_seq: 2-D list of probability distributions over each time - step, with each element being a list of normalized - probabilities over vocabulary and blank. - :type probs_seq: 2-D list - :param beam_size: Width for beam search. - :type beam_size: int - :param vocabulary: Vocabulary list. - :type vocabulary: list - :param cutoff_prob: Cutoff probability in pruning, - default 1.0, no pruning. - :type cutoff_prob: float - :param ext_scoring_func: External scoring function for - partially decoded sentence, e.g. word count - or language model. - :type external_scoring_func: callable - :param nproc: Whether the decoder used in multiprocesses. - :type nproc: bool - :return: List of tuples of log probability and sentence as decoding - results, in descending order of the probability. - :rtype: list - """ - # dimension check - for prob_list in probs_seq: - if not len(prob_list) == len(vocabulary) + 1: - raise ValueError("The shape of prob_seq does not match with the " - "shape of the vocabulary.") - - # blank_id assign - blank_id = len(vocabulary) - - # If the decoder called in the multiprocesses, then use the global scorer - # instantiated in ctc_beam_search_decoder_batch(). - if nproc is True: - global ext_nproc_scorer - ext_scoring_func = ext_nproc_scorer - - ## initialize - # prefix_set_prev: the set containing selected prefixes - # probs_b_prev: prefixes' probability ending with blank in previous step - # probs_nb_prev: prefixes' probability ending with non-blank in previous step - prefix_set_prev = {'\t': 1.0} - probs_b_prev, probs_nb_prev = {'\t': 1.0}, {'\t': 0.0} - - ## extend prefix in loop - for time_step in xrange(len(probs_seq)): - # prefix_set_next: the set containing candidate prefixes - # probs_b_cur: prefixes' probability ending with blank in current step - # probs_nb_cur: prefixes' probability ending with non-blank in current step - prefix_set_next, probs_b_cur, probs_nb_cur = {}, {}, {} - - prob_idx = list(enumerate(probs_seq[time_step])) - cutoff_len = len(prob_idx) - #If pruning is enabled - if cutoff_prob < 1.0 or cutoff_top_n < cutoff_len: - prob_idx = sorted(prob_idx, key=lambda asd: asd[1], reverse=True) - cutoff_len, cum_prob = 0, 0.0 - for i in xrange(len(prob_idx)): - cum_prob += prob_idx[i][1] - cutoff_len += 1 - if cum_prob >= cutoff_prob: - break - cutoff_len = min(cutoff_len, cutoff_top_n) - prob_idx = prob_idx[0:cutoff_len] - - for l in prefix_set_prev: - if not prefix_set_next.has_key(l): - probs_b_cur[l], probs_nb_cur[l] = 0.0, 0.0 - - # extend prefix by travering prob_idx - for index in xrange(cutoff_len): - c, prob_c = prob_idx[index][0], prob_idx[index][1] - - if c == blank_id: - probs_b_cur[l] += prob_c * ( - probs_b_prev[l] + probs_nb_prev[l]) - else: - last_char = l[-1] - new_char = vocabulary[c] - l_plus = l + new_char - if not prefix_set_next.has_key(l_plus): - probs_b_cur[l_plus], probs_nb_cur[l_plus] = 0.0, 0.0 - - if new_char == last_char: - probs_nb_cur[l_plus] += prob_c * probs_b_prev[l] - probs_nb_cur[l] += prob_c * probs_nb_prev[l] - elif new_char == ' ': - if (ext_scoring_func is None) or (len(l) == 1): - score = 1.0 - else: - prefix = l[1:] - score = ext_scoring_func(prefix) - probs_nb_cur[l_plus] += score * prob_c * ( - probs_b_prev[l] + probs_nb_prev[l]) - else: - probs_nb_cur[l_plus] += prob_c * ( - probs_b_prev[l] + probs_nb_prev[l]) - # add l_plus into prefix_set_next - prefix_set_next[l_plus] = probs_nb_cur[ - l_plus] + probs_b_cur[l_plus] - # add l into prefix_set_next - prefix_set_next[l] = probs_b_cur[l] + probs_nb_cur[l] - # update probs - probs_b_prev, probs_nb_prev = probs_b_cur, probs_nb_cur - - ## store top beam_size prefixes - prefix_set_prev = sorted( - prefix_set_next.iteritems(), key=lambda asd: asd[1], reverse=True) - if beam_size < len(prefix_set_prev): - prefix_set_prev = prefix_set_prev[:beam_size] - prefix_set_prev = dict(prefix_set_prev) - - beam_result = [] - for seq, prob in prefix_set_prev.items(): - if prob > 0.0 and len(seq) > 1: - result = seq[1:] - # score last word by external scorer - if (ext_scoring_func is not None) and (result[-1] != ' '): - prob = prob * ext_scoring_func(result) - log_prob = log(prob) - beam_result.append((log_prob, result)) - else: - beam_result.append((float('-inf'), '')) - - ## output top beam_size decoding results - beam_result = sorted(beam_result, key=lambda asd: asd[0], reverse=True) - return beam_result - - -def ctc_beam_search_decoder_batch(probs_split, - beam_size, - vocabulary, - num_processes, - cutoff_prob=1.0, - cutoff_top_n=40, - ext_scoring_func=None): - """CTC beam search decoder using multiple processes. - - :param probs_seq: 3-D list with each element as an instance of 2-D list - of probabilities used by ctc_beam_search_decoder(). - :type probs_seq: 3-D list - :param beam_size: Width for beam search. - :type beam_size: int - :param vocabulary: Vocabulary list. - :type vocabulary: list - :param num_processes: Number of parallel processes. - :type num_processes: int - :param cutoff_prob: Cutoff probability in pruning, - default 1.0, no pruning. - :type cutoff_prob: float - :param num_processes: Number of parallel processes. - :type num_processes: int - :param ext_scoring_func: External scoring function for - partially decoded sentence, e.g. word count - or language model. - :type external_scoring_function: callable - :return: List of tuples of log probability and sentence as decoding - results, in descending order of the probability. - :rtype: list - """ - if not num_processes > 0: - raise ValueError("Number of processes must be positive!") - - # use global variable to pass the externnal scorer to beam search decoder - global ext_nproc_scorer - ext_nproc_scorer = ext_scoring_func - nproc = True - - pool = multiprocessing.Pool(processes=num_processes) - results = [] - for i, probs_list in enumerate(probs_split): - args = (probs_list, beam_size, vocabulary, cutoff_prob, cutoff_top_n, - None, nproc) - results.append(pool.apply_async(ctc_beam_search_decoder, args)) - - pool.close() - pool.join() - beam_search_results = [result.get() for result in results] - return beam_search_results diff --git a/deep_speech_2/decoders/scorer_deprecated.py b/deep_speech_2/decoders/scorer_deprecated.py deleted file mode 100644 index c6a661030d4363727e259da9c7949e59705d55c8..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/scorer_deprecated.py +++ /dev/null @@ -1,68 +0,0 @@ -"""External Scorer for Beam Search Decoder.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os -import kenlm -import numpy as np - - -class Scorer(object): - """External scorer to evaluate a prefix or whole sentence in - beam search decoding, including the score from n-gram language - model and word count. - - :param alpha: Parameter associated with language model. Don't use - language model when alpha = 0. - :type alpha: float - :param beta: Parameter associated with word count. Don't use word - count when beta = 0. - :type beta: float - :model_path: Path to load language model. - :type model_path: basestring - """ - - def __init__(self, alpha, beta, model_path): - self._alpha = alpha - self._beta = beta - if not os.path.isfile(model_path): - raise IOError("Invaid language model path: %s" % model_path) - self._language_model = kenlm.LanguageModel(model_path) - - # n-gram language model scoring - def _language_model_score(self, sentence): - #log10 prob of last word - log_cond_prob = list( - self._language_model.full_scores(sentence, eos=False))[-1][0] - return np.power(10, log_cond_prob) - - # word insertion term - def _word_count(self, sentence): - words = sentence.strip().split(' ') - return len(words) - - # reset alpha and beta - def reset_params(self, alpha, beta): - self._alpha = alpha - self._beta = beta - - # execute evaluation - def __call__(self, sentence, log=False): - """Evaluation function, gathering all the different scores - and return the final one. - - :param sentence: The input sentence for evalutation - :type sentence: basestring - :param log: Whether return the score in log representation. - :type log: bool - :return: Evaluation score, in the decimal or log. - :rtype: float - """ - lm = self._language_model_score(sentence) - word_cnt = self._word_count(sentence) - if log == False: - score = np.power(lm, self._alpha) * np.power(word_cnt, self._beta) - else: - score = self._alpha * np.log(lm) + self._beta * np.log(word_cnt) - return score diff --git a/deep_speech_2/decoders/swig/__init__.py b/deep_speech_2/decoders/swig/__init__.py deleted file mode 100644 index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0000000000000000000000000000000000000000 diff --git a/deep_speech_2/decoders/swig/_init_paths.py b/deep_speech_2/decoders/swig/_init_paths.py deleted file mode 100644 index ddabb535be682d95c3c8b73003ea30eed06ca0b0..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/_init_paths.py +++ /dev/null @@ -1,19 +0,0 @@ -"""Set up paths for DS2""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os.path -import sys - - -def add_path(path): - if path not in sys.path: - sys.path.insert(0, path) - - -this_dir = os.path.dirname(__file__) - -# Add project path to PYTHONPATH -proj_path = os.path.join(this_dir, '..') -add_path(proj_path) diff --git a/deep_speech_2/decoders/swig/ctc_beam_search_decoder.cpp b/deep_speech_2/decoders/swig/ctc_beam_search_decoder.cpp deleted file mode 100644 index 4a63af26af5e2da135f75386581fbe61f56c7fc7..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/ctc_beam_search_decoder.cpp +++ /dev/null @@ -1,222 +0,0 @@ -#include "ctc_beam_search_decoder.h" - -#include -#include -#include -#include -#include -#include - -#include "ThreadPool.h" -#include "fst/fstlib.h" - -#include "decoder_utils.h" -#include "path_trie.h" - -using FSTMATCH = fst::SortedMatcher; - -std::vector> ctc_beam_search_decoder( - const std::vector> &probs_seq, - const std::vector &vocabulary, - size_t beam_size, - double cutoff_prob, - size_t cutoff_top_n, - Scorer *ext_scorer) { - // dimension check - size_t num_time_steps = probs_seq.size(); - for (size_t i = 0; i < num_time_steps; ++i) { - VALID_CHECK_EQ(probs_seq[i].size(), - vocabulary.size() + 1, - "The shape of probs_seq does not match with " - "the shape of the vocabulary"); - } - - // assign blank id - size_t blank_id = vocabulary.size(); - - // assign space id - auto it = std::find(vocabulary.begin(), vocabulary.end(), " "); - int space_id = it - vocabulary.begin(); - // if no space in vocabulary - if ((size_t)space_id >= vocabulary.size()) { - space_id = -2; - } - - // init prefixes' root - PathTrie root; - root.score = root.log_prob_b_prev = 0.0; - std::vector prefixes; - prefixes.push_back(&root); - - if (ext_scorer != nullptr && !ext_scorer->is_character_based()) { - auto fst_dict = static_cast(ext_scorer->dictionary); - fst::StdVectorFst *dict_ptr = fst_dict->Copy(true); - root.set_dictionary(dict_ptr); - auto matcher = std::make_shared(*dict_ptr, fst::MATCH_INPUT); - root.set_matcher(matcher); - } - - // prefix search over time - for (size_t time_step = 0; time_step < num_time_steps; ++time_step) { - auto &prob = probs_seq[time_step]; - - float min_cutoff = -NUM_FLT_INF; - bool full_beam = false; - if (ext_scorer != nullptr) { - size_t num_prefixes = std::min(prefixes.size(), beam_size); - std::sort( - prefixes.begin(), prefixes.begin() + num_prefixes, prefix_compare); - min_cutoff = prefixes[num_prefixes - 1]->score + - std::log(prob[blank_id]) - std::max(0.0, ext_scorer->beta); - full_beam = (num_prefixes == beam_size); - } - - std::vector> log_prob_idx = - get_pruned_log_probs(prob, cutoff_prob, cutoff_top_n); - // loop over chars - for (size_t index = 0; index < log_prob_idx.size(); index++) { - auto c = log_prob_idx[index].first; - auto log_prob_c = log_prob_idx[index].second; - - for (size_t i = 0; i < prefixes.size() && i < beam_size; ++i) { - auto prefix = prefixes[i]; - if (full_beam && log_prob_c + prefix->score < min_cutoff) { - break; - } - // blank - if (c == blank_id) { - prefix->log_prob_b_cur = - log_sum_exp(prefix->log_prob_b_cur, log_prob_c + prefix->score); - continue; - } - // repeated character - if (c == prefix->character) { - prefix->log_prob_nb_cur = log_sum_exp( - prefix->log_prob_nb_cur, log_prob_c + prefix->log_prob_nb_prev); - } - // get new prefix - auto prefix_new = prefix->get_path_trie(c); - - if (prefix_new != nullptr) { - float log_p = -NUM_FLT_INF; - - if (c == prefix->character && - prefix->log_prob_b_prev > -NUM_FLT_INF) { - log_p = log_prob_c + prefix->log_prob_b_prev; - } else if (c != prefix->character) { - log_p = log_prob_c + prefix->score; - } - - // language model scoring - if (ext_scorer != nullptr && - (c == space_id || ext_scorer->is_character_based())) { - PathTrie *prefix_to_score = nullptr; - // skip scoring the space - if (ext_scorer->is_character_based()) { - prefix_to_score = prefix_new; - } else { - prefix_to_score = prefix; - } - - float score = 0.0; - std::vector ngram; - ngram = ext_scorer->make_ngram(prefix_to_score); - score = ext_scorer->get_log_cond_prob(ngram) * ext_scorer->alpha; - log_p += score; - log_p += ext_scorer->beta; - } - prefix_new->log_prob_nb_cur = - log_sum_exp(prefix_new->log_prob_nb_cur, log_p); - } - } // end of loop over prefix - } // end of loop over vocabulary - - - prefixes.clear(); - // update log probs - root.iterate_to_vec(prefixes); - - // only preserve top beam_size prefixes - if (prefixes.size() >= beam_size) { - std::nth_element(prefixes.begin(), - prefixes.begin() + beam_size, - prefixes.end(), - prefix_compare); - for (size_t i = beam_size; i < prefixes.size(); ++i) { - prefixes[i]->remove(); - } - } - } // end of loop over time - - // score the last word of each prefix that doesn't end with space - if (ext_scorer != nullptr && !ext_scorer->is_character_based()) { - for (size_t i = 0; i < beam_size && i < prefixes.size(); ++i) { - auto prefix = prefixes[i]; - if (!prefix->is_empty() && prefix->character != space_id) { - float score = 0.0; - std::vector ngram = ext_scorer->make_ngram(prefix); - score = ext_scorer->get_log_cond_prob(ngram) * ext_scorer->alpha; - score += ext_scorer->beta; - prefix->score += score; - } - } - } - - size_t num_prefixes = std::min(prefixes.size(), beam_size); - std::sort(prefixes.begin(), prefixes.begin() + num_prefixes, prefix_compare); - - // compute aproximate ctc score as the return score, without affecting the - // return order of decoding result. To delete when decoder gets stable. - for (size_t i = 0; i < beam_size && i < prefixes.size(); ++i) { - double approx_ctc = prefixes[i]->score; - if (ext_scorer != nullptr) { - std::vector output; - prefixes[i]->get_path_vec(output); - auto prefix_length = output.size(); - auto words = ext_scorer->split_labels(output); - // remove word insert - approx_ctc = approx_ctc - prefix_length * ext_scorer->beta; - // remove language model weight: - approx_ctc -= (ext_scorer->get_sent_log_prob(words)) * ext_scorer->alpha; - } - prefixes[i]->approx_ctc = approx_ctc; - } - - return get_beam_search_result(prefixes, vocabulary, beam_size); -} - - -std::vector>> -ctc_beam_search_decoder_batch( - const std::vector>> &probs_split, - const std::vector &vocabulary, - size_t beam_size, - size_t num_processes, - double cutoff_prob, - size_t cutoff_top_n, - Scorer *ext_scorer) { - VALID_CHECK_GT(num_processes, 0, "num_processes must be nonnegative!"); - // thread pool - ThreadPool pool(num_processes); - // number of samples - size_t batch_size = probs_split.size(); - - // enqueue the tasks of decoding - std::vector>>> res; - for (size_t i = 0; i < batch_size; ++i) { - res.emplace_back(pool.enqueue(ctc_beam_search_decoder, - probs_split[i], - vocabulary, - beam_size, - cutoff_prob, - cutoff_top_n, - ext_scorer)); - } - - // get decoding results - std::vector>> batch_results; - for (size_t i = 0; i < batch_size; ++i) { - batch_results.emplace_back(res[i].get()); - } - return batch_results; -} diff --git a/deep_speech_2/decoders/swig/ctc_beam_search_decoder.h b/deep_speech_2/decoders/swig/ctc_beam_search_decoder.h deleted file mode 100644 index 6fdd15517e38021c5835837a65e4c30fdab2128e..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/ctc_beam_search_decoder.h +++ /dev/null @@ -1,61 +0,0 @@ -#ifndef CTC_BEAM_SEARCH_DECODER_H_ -#define CTC_BEAM_SEARCH_DECODER_H_ - -#include -#include -#include - -#include "scorer.h" - -/* CTC Beam Search Decoder - - * Parameters: - * probs_seq: 2-D vector that each element is a vector of probabilities - * over vocabulary of one time step. - * vocabulary: A vector of vocabulary. - * beam_size: The width of beam search. - * cutoff_prob: Cutoff probability for pruning. - * cutoff_top_n: Cutoff number for pruning. - * ext_scorer: External scorer to evaluate a prefix, which consists of - * n-gram language model scoring and word insertion term. - * Default null, decoding the input sample without scorer. - * Return: - * A vector that each element is a pair of score and decoding result, - * in desending order. -*/ -std::vector> ctc_beam_search_decoder( - const std::vector> &probs_seq, - const std::vector &vocabulary, - size_t beam_size, - double cutoff_prob = 1.0, - size_t cutoff_top_n = 40, - Scorer *ext_scorer = nullptr); - -/* CTC Beam Search Decoder for batch data - - * Parameters: - * probs_seq: 3-D vector that each element is a 2-D vector that can be used - * by ctc_beam_search_decoder(). - * vocabulary: A vector of vocabulary. - * beam_size: The width of beam search. - * num_processes: Number of threads for beam search. - * cutoff_prob: Cutoff probability for pruning. - * cutoff_top_n: Cutoff number for pruning. - * ext_scorer: External scorer to evaluate a prefix, which consists of - * n-gram language model scoring and word insertion term. - * Default null, decoding the input sample without scorer. - * Return: - * A 2-D vector that each element is a vector of beam search decoding - * result for one audio sample. -*/ -std::vector>> -ctc_beam_search_decoder_batch( - const std::vector>> &probs_split, - const std::vector &vocabulary, - size_t beam_size, - size_t num_processes, - double cutoff_prob = 1.0, - size_t cutoff_top_n = 40, - Scorer *ext_scorer = nullptr); - -#endif // CTC_BEAM_SEARCH_DECODER_H_ diff --git a/deep_speech_2/decoders/swig/ctc_greedy_decoder.cpp b/deep_speech_2/decoders/swig/ctc_greedy_decoder.cpp deleted file mode 100644 index 03449d7391514bd267b396bab31da2e498425b47..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/ctc_greedy_decoder.cpp +++ /dev/null @@ -1,45 +0,0 @@ -#include "ctc_greedy_decoder.h" -#include "decoder_utils.h" - -std::string ctc_greedy_decoder( - const std::vector> &probs_seq, - const std::vector &vocabulary) { - // dimension check - size_t num_time_steps = probs_seq.size(); - for (size_t i = 0; i < num_time_steps; ++i) { - VALID_CHECK_EQ(probs_seq[i].size(), - vocabulary.size() + 1, - "The shape of probs_seq does not match with " - "the shape of the vocabulary"); - } - - size_t blank_id = vocabulary.size(); - - std::vector max_idx_vec(num_time_steps, 0); - std::vector idx_vec; - for (size_t i = 0; i < num_time_steps; ++i) { - double max_prob = 0.0; - size_t max_idx = 0; - const std::vector &probs_step = probs_seq[i]; - for (size_t j = 0; j < probs_step.size(); ++j) { - if (max_prob < probs_step[j]) { - max_idx = j; - max_prob = probs_step[j]; - } - } - // id with maximum probability in current time step - max_idx_vec[i] = max_idx; - // deduplicate - if ((i == 0) || ((i > 0) && max_idx_vec[i] != max_idx_vec[i - 1])) { - idx_vec.push_back(max_idx_vec[i]); - } - } - - std::string best_path_result; - for (size_t i = 0; i < idx_vec.size(); ++i) { - if (idx_vec[i] != blank_id) { - best_path_result += vocabulary[idx_vec[i]]; - } - } - return best_path_result; -} diff --git a/deep_speech_2/decoders/swig/ctc_greedy_decoder.h b/deep_speech_2/decoders/swig/ctc_greedy_decoder.h deleted file mode 100644 index 5e64f692e5062c4f8f1d4aa0c8c7b75eaa0a668a..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/ctc_greedy_decoder.h +++ /dev/null @@ -1,20 +0,0 @@ -#ifndef CTC_GREEDY_DECODER_H -#define CTC_GREEDY_DECODER_H - -#include -#include - -/* CTC Greedy (Best Path) Decoder - * - * Parameters: - * probs_seq: 2-D vector that each element is a vector of probabilities - * over vocabulary of one time step. - * vocabulary: A vector of vocabulary. - * Return: - * The decoding result in string - */ -std::string ctc_greedy_decoder( - const std::vector>& probs_seq, - const std::vector& vocabulary); - -#endif // CTC_GREEDY_DECODER_H diff --git a/deep_speech_2/decoders/swig/decoder_utils.cpp b/deep_speech_2/decoders/swig/decoder_utils.cpp deleted file mode 100644 index 70a1592889bc24e0af344fc123101e1b19ff6c15..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/decoder_utils.cpp +++ /dev/null @@ -1,176 +0,0 @@ -#include "decoder_utils.h" - -#include -#include -#include - -std::vector> get_pruned_log_probs( - const std::vector &prob_step, - double cutoff_prob, - size_t cutoff_top_n) { - std::vector> prob_idx; - for (size_t i = 0; i < prob_step.size(); ++i) { - prob_idx.push_back(std::pair(i, prob_step[i])); - } - // pruning of vacobulary - size_t cutoff_len = prob_step.size(); - if (cutoff_prob < 1.0 || cutoff_top_n < cutoff_len) { - std::sort( - prob_idx.begin(), prob_idx.end(), pair_comp_second_rev); - if (cutoff_prob < 1.0) { - double cum_prob = 0.0; - cutoff_len = 0; - for (size_t i = 0; i < prob_idx.size(); ++i) { - cum_prob += prob_idx[i].second; - cutoff_len += 1; - if (cum_prob >= cutoff_prob || cutoff_len >= cutoff_top_n) break; - } - } - prob_idx = std::vector>( - prob_idx.begin(), prob_idx.begin() + cutoff_len); - } - std::vector> log_prob_idx; - for (size_t i = 0; i < cutoff_len; ++i) { - log_prob_idx.push_back(std::pair( - prob_idx[i].first, log(prob_idx[i].second + NUM_FLT_MIN))); - } - return log_prob_idx; -} - - -std::vector> get_beam_search_result( - const std::vector &prefixes, - const std::vector &vocabulary, - size_t beam_size) { - // allow for the post processing - std::vector space_prefixes; - if (space_prefixes.empty()) { - for (size_t i = 0; i < beam_size && i < prefixes.size(); ++i) { - space_prefixes.push_back(prefixes[i]); - } - } - - std::sort(space_prefixes.begin(), space_prefixes.end(), prefix_compare); - std::vector> output_vecs; - for (size_t i = 0; i < beam_size && i < space_prefixes.size(); ++i) { - std::vector output; - space_prefixes[i]->get_path_vec(output); - // convert index to string - std::string output_str; - for (size_t j = 0; j < output.size(); j++) { - output_str += vocabulary[output[j]]; - } - std::pair output_pair(-space_prefixes[i]->approx_ctc, - output_str); - output_vecs.emplace_back(output_pair); - } - - return output_vecs; -} - -size_t get_utf8_str_len(const std::string &str) { - size_t str_len = 0; - for (char c : str) { - str_len += ((c & 0xc0) != 0x80); - } - return str_len; -} - -std::vector split_utf8_str(const std::string &str) { - std::vector result; - std::string out_str; - - for (char c : str) { - if ((c & 0xc0) != 0x80) // new UTF-8 character - { - if (!out_str.empty()) { - result.push_back(out_str); - out_str.clear(); - } - } - - out_str.append(1, c); - } - result.push_back(out_str); - return result; -} - -std::vector split_str(const std::string &s, - const std::string &delim) { - std::vector result; - std::size_t start = 0, delim_len = delim.size(); - while (true) { - std::size_t end = s.find(delim, start); - if (end == std::string::npos) { - if (start < s.size()) { - result.push_back(s.substr(start)); - } - break; - } - if (end > start) { - result.push_back(s.substr(start, end - start)); - } - start = end + delim_len; - } - return result; -} - -bool prefix_compare(const PathTrie *x, const PathTrie *y) { - if (x->score == y->score) { - if (x->character == y->character) { - return false; - } else { - return (x->character < y->character); - } - } else { - return x->score > y->score; - } -} - -void add_word_to_fst(const std::vector &word, - fst::StdVectorFst *dictionary) { - if (dictionary->NumStates() == 0) { - fst::StdVectorFst::StateId start = dictionary->AddState(); - assert(start == 0); - dictionary->SetStart(start); - } - fst::StdVectorFst::StateId src = dictionary->Start(); - fst::StdVectorFst::StateId dst; - for (auto c : word) { - dst = dictionary->AddState(); - dictionary->AddArc(src, fst::StdArc(c, c, 0, dst)); - src = dst; - } - dictionary->SetFinal(dst, fst::StdArc::Weight::One()); -} - -bool add_word_to_dictionary( - const std::string &word, - const std::unordered_map &char_map, - bool add_space, - int SPACE_ID, - fst::StdVectorFst *dictionary) { - auto characters = split_utf8_str(word); - - std::vector int_word; - - for (auto &c : characters) { - if (c == " ") { - int_word.push_back(SPACE_ID); - } else { - auto int_c = char_map.find(c); - if (int_c != char_map.end()) { - int_word.push_back(int_c->second); - } else { - return false; // return without adding - } - } - } - - if (add_space) { - int_word.push_back(SPACE_ID); - } - - add_word_to_fst(int_word, dictionary); - return true; // return with successful adding -} diff --git a/deep_speech_2/decoders/swig/decoder_utils.h b/deep_speech_2/decoders/swig/decoder_utils.h deleted file mode 100644 index 72821c187fff9567645de82b1c45fc0350787173..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/decoder_utils.h +++ /dev/null @@ -1,94 +0,0 @@ -#ifndef DECODER_UTILS_H_ -#define DECODER_UTILS_H_ - -#include -#include "fst/log.h" -#include "path_trie.h" - -const float NUM_FLT_INF = std::numeric_limits::max(); -const float NUM_FLT_MIN = std::numeric_limits::min(); - -// inline function for validation check -inline void check( - bool x, const char *expr, const char *file, int line, const char *err) { - if (!x) { - std::cout << "[" << file << ":" << line << "] "; - LOG(FATAL) << "\"" << expr << "\" check failed. " << err; - } -} - -#define VALID_CHECK(x, info) \ - check(static_cast(x), #x, __FILE__, __LINE__, info) -#define VALID_CHECK_EQ(x, y, info) VALID_CHECK((x) == (y), info) -#define VALID_CHECK_GT(x, y, info) VALID_CHECK((x) > (y), info) -#define VALID_CHECK_LT(x, y, info) VALID_CHECK((x) < (y), info) - - -// Function template for comparing two pairs -template -bool pair_comp_first_rev(const std::pair &a, - const std::pair &b) { - return a.first > b.first; -} - -// Function template for comparing two pairs -template -bool pair_comp_second_rev(const std::pair &a, - const std::pair &b) { - return a.second > b.second; -} - -// Return the sum of two probabilities in log scale -template -T log_sum_exp(const T &x, const T &y) { - static T num_min = -std::numeric_limits::max(); - if (x <= num_min) return y; - if (y <= num_min) return x; - T xmax = std::max(x, y); - return std::log(std::exp(x - xmax) + std::exp(y - xmax)) + xmax; -} - -// Get pruned probability vector for each time step's beam search -std::vector> get_pruned_log_probs( - const std::vector &prob_step, - double cutoff_prob, - size_t cutoff_top_n); - -// Get beam search result from prefixes in trie tree -std::vector> get_beam_search_result( - const std::vector &prefixes, - const std::vector &vocabulary, - size_t beam_size); - -// Functor for prefix comparsion -bool prefix_compare(const PathTrie *x, const PathTrie *y); - -/* Get length of utf8 encoding string - * See: http://stackoverflow.com/a/4063229 - */ -size_t get_utf8_str_len(const std::string &str); - -/* Split a string into a list of strings on a given string - * delimiter. NB: delimiters on beginning / end of string are - * trimmed. Eg, "FooBarFoo" split on "Foo" returns ["Bar"]. - */ -std::vector split_str(const std::string &s, - const std::string &delim); - -/* Splits string into vector of strings representing - * UTF-8 characters (not same as chars) - */ -std::vector split_utf8_str(const std::string &str); - -// Add a word in index to the dicionary of fst -void add_word_to_fst(const std::vector &word, - fst::StdVectorFst *dictionary); - -// Add a word in string to dictionary -bool add_word_to_dictionary( - const std::string &word, - const std::unordered_map &char_map, - bool add_space, - int SPACE_ID, - fst::StdVectorFst *dictionary); -#endif // DECODER_UTILS_H diff --git a/deep_speech_2/decoders/swig/decoders.i b/deep_speech_2/decoders/swig/decoders.i deleted file mode 100644 index 4227d4a37528e30fa964e107b430896038c07dcf..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/decoders.i +++ /dev/null @@ -1,33 +0,0 @@ -%module swig_decoders -%{ -#include "scorer.h" -#include "ctc_greedy_decoder.h" -#include "ctc_beam_search_decoder.h" -#include "decoder_utils.h" -%} - -%include "std_vector.i" -%include "std_pair.i" -%include "std_string.i" -%import "decoder_utils.h" - -namespace std { - %template(DoubleVector) std::vector; - %template(IntVector) std::vector; - %template(StringVector) std::vector; - %template(VectorOfStructVector) std::vector >; - %template(FloatVector) std::vector; - %template(Pair) std::pair; - %template(PairFloatStringVector) std::vector >; - %template(PairDoubleStringVector) std::vector >; - %template(PairDoubleStringVector2) std::vector > >; - %template(DoubleVector3) std::vector > >; -} - -%template(IntDoublePairCompSecondRev) pair_comp_second_rev; -%template(StringDoublePairCompSecondRev) pair_comp_second_rev; -%template(DoubleStringPairCompFirstRev) pair_comp_first_rev; - -%include "scorer.h" -%include "ctc_greedy_decoder.h" -%include "ctc_beam_search_decoder.h" diff --git a/deep_speech_2/decoders/swig/path_trie.cpp b/deep_speech_2/decoders/swig/path_trie.cpp deleted file mode 100644 index 40d9097055686eb9de0834fa2ab9478988bbfa96..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/path_trie.cpp +++ /dev/null @@ -1,148 +0,0 @@ -#include "path_trie.h" - -#include -#include -#include -#include -#include - -#include "decoder_utils.h" - -PathTrie::PathTrie() { - log_prob_b_prev = -NUM_FLT_INF; - log_prob_nb_prev = -NUM_FLT_INF; - log_prob_b_cur = -NUM_FLT_INF; - log_prob_nb_cur = -NUM_FLT_INF; - score = -NUM_FLT_INF; - - ROOT_ = -1; - character = ROOT_; - exists_ = true; - parent = nullptr; - - dictionary_ = nullptr; - dictionary_state_ = 0; - has_dictionary_ = false; - - matcher_ = nullptr; -} - -PathTrie::~PathTrie() { - for (auto child : children_) { - delete child.second; - } -} - -PathTrie* PathTrie::get_path_trie(int new_char, bool reset) { - auto child = children_.begin(); - for (child = children_.begin(); child != children_.end(); ++child) { - if (child->first == new_char) { - break; - } - } - if (child != children_.end()) { - if (!child->second->exists_) { - child->second->exists_ = true; - child->second->log_prob_b_prev = -NUM_FLT_INF; - child->second->log_prob_nb_prev = -NUM_FLT_INF; - child->second->log_prob_b_cur = -NUM_FLT_INF; - child->second->log_prob_nb_cur = -NUM_FLT_INF; - } - return (child->second); - } else { - if (has_dictionary_) { - matcher_->SetState(dictionary_state_); - bool found = matcher_->Find(new_char); - if (!found) { - // Adding this character causes word outside dictionary - auto FSTZERO = fst::TropicalWeight::Zero(); - auto final_weight = dictionary_->Final(dictionary_state_); - bool is_final = (final_weight != FSTZERO); - if (is_final && reset) { - dictionary_state_ = dictionary_->Start(); - } - return nullptr; - } else { - PathTrie* new_path = new PathTrie; - new_path->character = new_char; - new_path->parent = this; - new_path->dictionary_ = dictionary_; - new_path->dictionary_state_ = matcher_->Value().nextstate; - new_path->has_dictionary_ = true; - new_path->matcher_ = matcher_; - children_.push_back(std::make_pair(new_char, new_path)); - return new_path; - } - } else { - PathTrie* new_path = new PathTrie; - new_path->character = new_char; - new_path->parent = this; - children_.push_back(std::make_pair(new_char, new_path)); - return new_path; - } - } -} - -PathTrie* PathTrie::get_path_vec(std::vector& output) { - return get_path_vec(output, ROOT_); -} - -PathTrie* PathTrie::get_path_vec(std::vector& output, - int stop, - size_t max_steps) { - if (character == stop || character == ROOT_ || output.size() == max_steps) { - std::reverse(output.begin(), output.end()); - return this; - } else { - output.push_back(character); - return parent->get_path_vec(output, stop, max_steps); - } -} - -void PathTrie::iterate_to_vec(std::vector& output) { - if (exists_) { - log_prob_b_prev = log_prob_b_cur; - log_prob_nb_prev = log_prob_nb_cur; - - log_prob_b_cur = -NUM_FLT_INF; - log_prob_nb_cur = -NUM_FLT_INF; - - score = log_sum_exp(log_prob_b_prev, log_prob_nb_prev); - output.push_back(this); - } - for (auto child : children_) { - child.second->iterate_to_vec(output); - } -} - -void PathTrie::remove() { - exists_ = false; - - if (children_.size() == 0) { - auto child = parent->children_.begin(); - for (child = parent->children_.begin(); child != parent->children_.end(); - ++child) { - if (child->first == character) { - parent->children_.erase(child); - break; - } - } - - if (parent->children_.size() == 0 && !parent->exists_) { - parent->remove(); - } - - delete this; - } -} - -void PathTrie::set_dictionary(fst::StdVectorFst* dictionary) { - dictionary_ = dictionary; - dictionary_state_ = dictionary->Start(); - has_dictionary_ = true; -} - -using FSTMATCH = fst::SortedMatcher; -void PathTrie::set_matcher(std::shared_ptr matcher) { - matcher_ = matcher; -} diff --git a/deep_speech_2/decoders/swig/path_trie.h b/deep_speech_2/decoders/swig/path_trie.h deleted file mode 100644 index 7fd715d26d74a8299d371e17f87d6d309a39289b..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/path_trie.h +++ /dev/null @@ -1,67 +0,0 @@ -#ifndef PATH_TRIE_H -#define PATH_TRIE_H - -#include -#include -#include -#include -#include - -#include "fst/fstlib.h" - -/* Trie tree for prefix storing and manipulating, with a dictionary in - * finite-state transducer for spelling correction. - */ -class PathTrie { -public: - PathTrie(); - ~PathTrie(); - - // get new prefix after appending new char - PathTrie* get_path_trie(int new_char, bool reset = true); - - // get the prefix in index from root to current node - PathTrie* get_path_vec(std::vector& output); - - // get the prefix in index from some stop node to current nodel - PathTrie* get_path_vec(std::vector& output, - int stop, - size_t max_steps = std::numeric_limits::max()); - - // update log probs - void iterate_to_vec(std::vector& output); - - // set dictionary for FST - void set_dictionary(fst::StdVectorFst* dictionary); - - void set_matcher(std::shared_ptr>); - - bool is_empty() { return ROOT_ == character; } - - // remove current path from root - void remove(); - - float log_prob_b_prev; - float log_prob_nb_prev; - float log_prob_b_cur; - float log_prob_nb_cur; - float score; - float approx_ctc; - int character; - PathTrie* parent; - -private: - int ROOT_; - bool exists_; - bool has_dictionary_; - - std::vector> children_; - - // pointer to dictionary of FST - fst::StdVectorFst* dictionary_; - fst::StdVectorFst::StateId dictionary_state_; - // true if finding ars in FST - std::shared_ptr> matcher_; -}; - -#endif // PATH_TRIE_H diff --git a/deep_speech_2/decoders/swig/scorer.cpp b/deep_speech_2/decoders/swig/scorer.cpp deleted file mode 100644 index 686c67c77e1d7d1267df963308582bf194859310..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/scorer.cpp +++ /dev/null @@ -1,234 +0,0 @@ -#include "scorer.h" - -#include -#include - -#include "lm/config.hh" -#include "lm/model.hh" -#include "lm/state.hh" -#include "util/string_piece.hh" -#include "util/tokenize_piece.hh" - -#include "decoder_utils.h" - -using namespace lm::ngram; - -Scorer::Scorer(double alpha, - double beta, - const std::string& lm_path, - const std::vector& vocab_list) { - this->alpha = alpha; - this->beta = beta; - - dictionary = nullptr; - is_character_based_ = true; - language_model_ = nullptr; - - max_order_ = 0; - dict_size_ = 0; - SPACE_ID_ = -1; - - setup(lm_path, vocab_list); -} - -Scorer::~Scorer() { - if (language_model_ != nullptr) { - delete static_cast(language_model_); - } - if (dictionary != nullptr) { - delete static_cast(dictionary); - } -} - -void Scorer::setup(const std::string& lm_path, - const std::vector& vocab_list) { - // load language model - load_lm(lm_path); - // set char map for scorer - set_char_map(vocab_list); - // fill the dictionary for FST - if (!is_character_based()) { - fill_dictionary(true); - } -} - -void Scorer::load_lm(const std::string& lm_path) { - const char* filename = lm_path.c_str(); - VALID_CHECK_EQ(access(filename, F_OK), 0, "Invalid language model path"); - - RetriveStrEnumerateVocab enumerate; - lm::ngram::Config config; - config.enumerate_vocab = &enumerate; - language_model_ = lm::ngram::LoadVirtual(filename, config); - max_order_ = static_cast(language_model_)->Order(); - vocabulary_ = enumerate.vocabulary; - for (size_t i = 0; i < vocabulary_.size(); ++i) { - if (is_character_based_ && vocabulary_[i] != UNK_TOKEN && - vocabulary_[i] != START_TOKEN && vocabulary_[i] != END_TOKEN && - get_utf8_str_len(enumerate.vocabulary[i]) > 1) { - is_character_based_ = false; - } - } -} - -double Scorer::get_log_cond_prob(const std::vector& words) { - lm::base::Model* model = static_cast(language_model_); - double cond_prob; - lm::ngram::State state, tmp_state, out_state; - // avoid to inserting in begin - model->NullContextWrite(&state); - for (size_t i = 0; i < words.size(); ++i) { - lm::WordIndex word_index = model->BaseVocabulary().Index(words[i]); - // encounter OOV - if (word_index == 0) { - return OOV_SCORE; - } - cond_prob = model->BaseScore(&state, word_index, &out_state); - tmp_state = state; - state = out_state; - out_state = tmp_state; - } - // return log10 prob - return cond_prob; -} - -double Scorer::get_sent_log_prob(const std::vector& words) { - std::vector sentence; - if (words.size() == 0) { - for (size_t i = 0; i < max_order_; ++i) { - sentence.push_back(START_TOKEN); - } - } else { - for (size_t i = 0; i < max_order_ - 1; ++i) { - sentence.push_back(START_TOKEN); - } - sentence.insert(sentence.end(), words.begin(), words.end()); - } - sentence.push_back(END_TOKEN); - return get_log_prob(sentence); -} - -double Scorer::get_log_prob(const std::vector& words) { - assert(words.size() > max_order_); - double score = 0.0; - for (size_t i = 0; i < words.size() - max_order_ + 1; ++i) { - std::vector ngram(words.begin() + i, - words.begin() + i + max_order_); - score += get_log_cond_prob(ngram); - } - return score; -} - -void Scorer::reset_params(float alpha, float beta) { - this->alpha = alpha; - this->beta = beta; -} - -std::string Scorer::vec2str(const std::vector& input) { - std::string word; - for (auto ind : input) { - word += char_list_[ind]; - } - return word; -} - -std::vector Scorer::split_labels(const std::vector& labels) { - if (labels.empty()) return {}; - - std::string s = vec2str(labels); - std::vector words; - if (is_character_based_) { - words = split_utf8_str(s); - } else { - words = split_str(s, " "); - } - return words; -} - -void Scorer::set_char_map(const std::vector& char_list) { - char_list_ = char_list; - char_map_.clear(); - - for (size_t i = 0; i < char_list_.size(); i++) { - if (char_list_[i] == " ") { - SPACE_ID_ = i; - char_map_[' '] = i; - } else if (char_list_[i].size() == 1) { - char_map_[char_list_[i][0]] = i; - } - } -} - -std::vector Scorer::make_ngram(PathTrie* prefix) { - std::vector ngram; - PathTrie* current_node = prefix; - PathTrie* new_node = nullptr; - - for (int order = 0; order < max_order_; order++) { - std::vector prefix_vec; - - if (is_character_based_) { - new_node = current_node->get_path_vec(prefix_vec, SPACE_ID_, 1); - current_node = new_node; - } else { - new_node = current_node->get_path_vec(prefix_vec, SPACE_ID_); - current_node = new_node->parent; // Skipping spaces - } - - // reconstruct word - std::string word = vec2str(prefix_vec); - ngram.push_back(word); - - if (new_node->character == -1) { - // No more spaces, but still need order - for (int i = 0; i < max_order_ - order - 1; i++) { - ngram.push_back(START_TOKEN); - } - break; - } - } - std::reverse(ngram.begin(), ngram.end()); - return ngram; -} - -void Scorer::fill_dictionary(bool add_space) { - fst::StdVectorFst dictionary; - // First reverse char_list so ints can be accessed by chars - std::unordered_map char_map; - for (size_t i = 0; i < char_list_.size(); i++) { - char_map[char_list_[i]] = i; - } - - // For each unigram convert to ints and put in trie - int dict_size = 0; - for (const auto& word : vocabulary_) { - bool added = add_word_to_dictionary( - word, char_map, add_space, SPACE_ID_, &dictionary); - dict_size += added ? 1 : 0; - } - - dict_size_ = dict_size; - - /* Simplify FST - - * This gets rid of "epsilon" transitions in the FST. - * These are transitions that don't require a string input to be taken. - * Getting rid of them is necessary to make the FST determinisitc, but - * can greatly increase the size of the FST - */ - fst::RmEpsilon(&dictionary); - fst::StdVectorFst* new_dict = new fst::StdVectorFst; - - /* This makes the FST deterministic, meaning for any string input there's - * only one possible state the FST could be in. It is assumed our - * dictionary is deterministic when using it. - * (lest we'd have to check for multiple transitions at each state) - */ - fst::Determinize(dictionary, new_dict); - - /* Finds the simplest equivalent fst. This is unnecessary but decreases - * memory usage of the dictionary - */ - fst::Minimize(new_dict); - this->dictionary = new_dict; -} diff --git a/deep_speech_2/decoders/swig/scorer.h b/deep_speech_2/decoders/swig/scorer.h deleted file mode 100644 index 618364635976db967f0c7e89bf4aa22dad63ff46..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/scorer.h +++ /dev/null @@ -1,112 +0,0 @@ -#ifndef SCORER_H_ -#define SCORER_H_ - -#include -#include -#include -#include - -#include "lm/enumerate_vocab.hh" -#include "lm/virtual_interface.hh" -#include "lm/word_index.hh" -#include "util/string_piece.hh" - -#include "path_trie.h" - -const double OOV_SCORE = -1000.0; -const std::string START_TOKEN = ""; -const std::string UNK_TOKEN = ""; -const std::string END_TOKEN = ""; - -// Implement a callback to retrive the dictionary of language model. -class RetriveStrEnumerateVocab : public lm::EnumerateVocab { -public: - RetriveStrEnumerateVocab() {} - - void Add(lm::WordIndex index, const StringPiece &str) { - vocabulary.push_back(std::string(str.data(), str.length())); - } - - std::vector vocabulary; -}; - -/* External scorer to query score for n-gram or sentence, including language - * model scoring and word insertion. - * - * Example: - * Scorer scorer(alpha, beta, "path_of_language_model"); - * scorer.get_log_cond_prob({ "WORD1", "WORD2", "WORD3" }); - * scorer.get_sent_log_prob({ "WORD1", "WORD2", "WORD3" }); - */ -class Scorer { -public: - Scorer(double alpha, - double beta, - const std::string &lm_path, - const std::vector &vocabulary); - ~Scorer(); - - double get_log_cond_prob(const std::vector &words); - - double get_sent_log_prob(const std::vector &words); - - // return the max order - size_t get_max_order() const { return max_order_; } - - // return the dictionary size of language model - size_t get_dict_size() const { return dict_size_; } - - // retrun true if the language model is character based - bool is_character_based() const { return is_character_based_; } - - // reset params alpha & beta - void reset_params(float alpha, float beta); - - // make ngram for a given prefix - std::vector make_ngram(PathTrie *prefix); - - // trransform the labels in index to the vector of words (word based lm) or - // the vector of characters (character based lm) - std::vector split_labels(const std::vector &labels); - - // language model weight - double alpha; - // word insertion weight - double beta; - - // pointer to the dictionary of FST - void *dictionary; - -protected: - // necessary setup: load language model, set char map, fill FST's dictionary - void setup(const std::string &lm_path, - const std::vector &vocab_list); - - // load language model from given path - void load_lm(const std::string &lm_path); - - // fill dictionary for FST - void fill_dictionary(bool add_space); - - // set char map - void set_char_map(const std::vector &char_list); - - double get_log_prob(const std::vector &words); - - // translate the vector in index to string - std::string vec2str(const std::vector &input); - -private: - void *language_model_; - bool is_character_based_; - size_t max_order_; - size_t dict_size_; - - int SPACE_ID_; - std::vector char_list_; - std::unordered_map char_map_; - - std::vector vocabulary_; -}; - -#endif // SCORER_H_ diff --git a/deep_speech_2/decoders/swig/setup.py b/deep_speech_2/decoders/swig/setup.py deleted file mode 100644 index b6bc0ca06af53315ac951dc793112c1cca7adcd4..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/setup.py +++ /dev/null @@ -1,119 +0,0 @@ -"""Script to build and install decoder package.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -from setuptools import setup, Extension, distutils -import glob -import platform -import os, sys -import multiprocessing.pool -import argparse - -parser = argparse.ArgumentParser(description=__doc__) -parser.add_argument( - "--num_processes", - default=1, - type=int, - help="Number of cpu processes to build package. (default: %(default)d)") -args = parser.parse_known_args() - -# reconstruct sys.argv to pass to setup below -sys.argv = [sys.argv[0]] + args[1] - - -# monkey-patch for parallel compilation -# See: https://stackoverflow.com/a/13176803 -def parallelCCompile(self, - sources, - output_dir=None, - macros=None, - include_dirs=None, - debug=0, - extra_preargs=None, - extra_postargs=None, - depends=None): - # those lines are copied from distutils.ccompiler.CCompiler directly - macros, objects, extra_postargs, pp_opts, build = self._setup_compile( - output_dir, macros, include_dirs, sources, depends, extra_postargs) - cc_args = self._get_cc_args(pp_opts, debug, extra_preargs) - - # parallel code - def _single_compile(obj): - try: - src, ext = build[obj] - except KeyError: - return - self._compile(obj, src, ext, cc_args, extra_postargs, pp_opts) - - # convert to list, imap is evaluated on-demand - thread_pool = multiprocessing.pool.ThreadPool(args[0].num_processes) - list(thread_pool.imap(_single_compile, objects)) - return objects - - -def compile_test(header, library): - dummy_path = os.path.join(os.path.dirname(__file__), "dummy") - command = "bash -c \"g++ -include " + header \ - + " -l" + library + " -x c++ - <<<'int main() {}' -o " \ - + dummy_path + " >/dev/null 2>/dev/null && rm " \ - + dummy_path + " 2>/dev/null\"" - return os.system(command) == 0 - - -# hack compile to support parallel compiling -distutils.ccompiler.CCompiler.compile = parallelCCompile - -FILES = glob.glob('kenlm/util/*.cc') \ - + glob.glob('kenlm/lm/*.cc') \ - + glob.glob('kenlm/util/double-conversion/*.cc') - -FILES += glob.glob('openfst-1.6.3/src/lib/*.cc') - -FILES = [ - fn for fn in FILES - if not (fn.endswith('main.cc') or fn.endswith('test.cc') or fn.endswith( - 'unittest.cc')) -] - -LIBS = ['stdc++'] -if platform.system() != 'Darwin': - LIBS.append('rt') - -ARGS = ['-O3', '-DNDEBUG', '-DKENLM_MAX_ORDER=6', '-std=c++11'] - -if compile_test('zlib.h', 'z'): - ARGS.append('-DHAVE_ZLIB') - LIBS.append('z') - -if compile_test('bzlib.h', 'bz2'): - ARGS.append('-DHAVE_BZLIB') - LIBS.append('bz2') - -if compile_test('lzma.h', 'lzma'): - ARGS.append('-DHAVE_XZLIB') - LIBS.append('lzma') - -os.system('swig -python -c++ ./decoders.i') - -decoders_module = [ - Extension( - name='_swig_decoders', - sources=FILES + glob.glob('*.cxx') + glob.glob('*.cpp'), - language='c++', - include_dirs=[ - '.', - 'kenlm', - 'openfst-1.6.3/src/include', - 'ThreadPool', - ], - libraries=LIBS, - extra_compile_args=ARGS) -] - -setup( - name='swig_decoders', - version='1.0', - description="""CTC decoders""", - ext_modules=decoders_module, - py_modules=['swig_decoders'], ) diff --git a/deep_speech_2/decoders/swig/setup.sh b/deep_speech_2/decoders/swig/setup.sh deleted file mode 100644 index 78ae2b2011c1a0296892358941280d9ffb702b1e..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig/setup.sh +++ /dev/null @@ -1,21 +0,0 @@ -#!/usr/bin/env bash - -if [ ! -d kenlm ]; then - git clone https://github.com/luotao1/kenlm.git - echo -e "\n" -fi - -if [ ! -d openfst-1.6.3 ]; then - echo "Download and extract openfst ..." - wget http://www.openfst.org/twiki/pub/FST/FstDownload/openfst-1.6.3.tar.gz - tar -xzvf openfst-1.6.3.tar.gz - echo -e "\n" -fi - -if [ ! -d ThreadPool ]; then - git clone https://github.com/progschj/ThreadPool.git - echo -e "\n" -fi - -echo "Install decoders ..." -python setup.py install --num_processes 4 diff --git a/deep_speech_2/decoders/swig_wrapper.py b/deep_speech_2/decoders/swig_wrapper.py deleted file mode 100644 index 21aed03c1b551383a75aa4019c4cb002e48254b2..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/swig_wrapper.py +++ /dev/null @@ -1,124 +0,0 @@ -"""Wrapper for various CTC decoders in SWIG.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import swig_decoders - - -class Scorer(swig_decoders.Scorer): - """Wrapper for Scorer. - - :param alpha: Parameter associated with language model. Don't use - language model when alpha = 0. - :type alpha: float - :param beta: Parameter associated with word count. Don't use word - count when beta = 0. - :type beta: float - :model_path: Path to load language model. - :type model_path: basestring - """ - - def __init__(self, alpha, beta, model_path, vocabulary): - swig_decoders.Scorer.__init__(self, alpha, beta, model_path, vocabulary) - - -def ctc_greedy_decoder(probs_seq, vocabulary): - """Wrapper for ctc best path decoder in swig. - - :param probs_seq: 2-D list of probability distributions over each time - step, with each element being a list of normalized - probabilities over vocabulary and blank. - :type probs_seq: 2-D list - :param vocabulary: Vocabulary list. - :type vocabulary: list - :return: Decoding result string. - :rtype: basestring - """ - result = swig_decoders.ctc_greedy_decoder(probs_seq.tolist(), vocabulary) - return result.decode('utf-8') - - -def ctc_beam_search_decoder(probs_seq, - vocabulary, - beam_size, - cutoff_prob=1.0, - cutoff_top_n=40, - ext_scoring_func=None): - """Wrapper for the CTC Beam Search Decoder. - - :param probs_seq: 2-D list of probability distributions over each time - step, with each element being a list of normalized - probabilities over vocabulary and blank. - :type probs_seq: 2-D list - :param vocabulary: Vocabulary list. - :type vocabulary: list - :param beam_size: Width for beam search. - :type beam_size: int - :param cutoff_prob: Cutoff probability in pruning, - default 1.0, no pruning. - :type cutoff_prob: float - :param cutoff_top_n: Cutoff number in pruning, only top cutoff_top_n - characters with highest probs in vocabulary will be - used in beam search, default 40. - :type cutoff_top_n: int - :param ext_scoring_func: External scoring function for - partially decoded sentence, e.g. word count - or language model. - :type external_scoring_func: callable - :return: List of tuples of log probability and sentence as decoding - results, in descending order of the probability. - :rtype: list - """ - beam_results = swig_decoders.ctc_beam_search_decoder( - probs_seq.tolist(), vocabulary, beam_size, cutoff_prob, cutoff_top_n, - ext_scoring_func) - beam_results = [(res[0], res[1].decode('utf-8')) for res in beam_results] - return beam_results - - -def ctc_beam_search_decoder_batch(probs_split, - vocabulary, - beam_size, - num_processes, - cutoff_prob=1.0, - cutoff_top_n=40, - ext_scoring_func=None): - """Wrapper for the batched CTC beam search decoder. - - :param probs_seq: 3-D list with each element as an instance of 2-D list - of probabilities used by ctc_beam_search_decoder(). - :type probs_seq: 3-D list - :param vocabulary: Vocabulary list. - :type vocabulary: list - :param beam_size: Width for beam search. - :type beam_size: int - :param num_processes: Number of parallel processes. - :type num_processes: int - :param cutoff_prob: Cutoff probability in vocabulary pruning, - default 1.0, no pruning. - :type cutoff_prob: float - :param cutoff_top_n: Cutoff number in pruning, only top cutoff_top_n - characters with highest probs in vocabulary will be - used in beam search, default 40. - :type cutoff_top_n: int - :param num_processes: Number of parallel processes. - :type num_processes: int - :param ext_scoring_func: External scoring function for - partially decoded sentence, e.g. word count - or language model. - :type external_scoring_function: callable - :return: List of tuples of log probability and sentence as decoding - results, in descending order of the probability. - :rtype: list - """ - probs_split = [probs_seq.tolist() for probs_seq in probs_split] - - batch_beam_results = swig_decoders.ctc_beam_search_decoder_batch( - probs_split, vocabulary, beam_size, num_processes, cutoff_prob, - cutoff_top_n, ext_scoring_func) - batch_beam_results = [ - [(res[0], res[1].decode("utf-8")) for res in beam_results] - for beam_results in batch_beam_results - ] - return batch_beam_results diff --git a/deep_speech_2/decoders/tests/test_decoders.py b/deep_speech_2/decoders/tests/test_decoders.py deleted file mode 100644 index d522b5efa0b1eb821b9b2fb7616b298af57f60a7..0000000000000000000000000000000000000000 --- a/deep_speech_2/decoders/tests/test_decoders.py +++ /dev/null @@ -1,90 +0,0 @@ -"""Test decoders.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import unittest -from decoders import decoders_deprecated as decoder - - -class TestDecoders(unittest.TestCase): - def setUp(self): - self.vocab_list = ["\'", ' ', 'a', 'b', 'c', 'd'] - self.beam_size = 20 - self.probs_seq1 = [[ - 0.06390443, 0.21124858, 0.27323887, 0.06870235, 0.0361254, - 0.18184413, 0.16493624 - ], [ - 0.03309247, 0.22866108, 0.24390638, 0.09699597, 0.31895462, - 0.0094893, 0.06890021 - ], [ - 0.218104, 0.19992557, 0.18245131, 0.08503348, 0.14903535, - 0.08424043, 0.08120984 - ], [ - 0.12094152, 0.19162472, 0.01473646, 0.28045061, 0.24246305, - 0.05206269, 0.09772094 - ], [ - 0.1333387, 0.00550838, 0.00301669, 0.21745861, 0.20803985, - 0.41317442, 0.01946335 - ], [ - 0.16468227, 0.1980699, 0.1906545, 0.18963251, 0.19860937, - 0.04377724, 0.01457421 - ]] - self.probs_seq2 = [[ - 0.08034842, 0.22671944, 0.05799633, 0.36814645, 0.11307441, - 0.04468023, 0.10903471 - ], [ - 0.09742457, 0.12959763, 0.09435383, 0.21889204, 0.15113123, - 0.10219457, 0.20640612 - ], [ - 0.45033529, 0.09091417, 0.15333208, 0.07939558, 0.08649316, - 0.12298585, 0.01654384 - ], [ - 0.02512238, 0.22079203, 0.19664364, 0.11906379, 0.07816055, - 0.22538587, 0.13483174 - ], [ - 0.17928453, 0.06065261, 0.41153005, 0.1172041, 0.11880313, - 0.07113197, 0.04139363 - ], [ - 0.15882358, 0.1235788, 0.23376776, 0.20510435, 0.00279306, - 0.05294827, 0.22298418 - ]] - self.greedy_result = ["ac'bdc", "b'da"] - self.beam_search_result = ['acdc', "b'a"] - - def test_greedy_decoder_1(self): - bst_result = decoder.ctc_greedy_decoder(self.probs_seq1, - self.vocab_list) - self.assertEqual(bst_result, self.greedy_result[0]) - - def test_greedy_decoder_2(self): - bst_result = decoder.ctc_greedy_decoder(self.probs_seq2, - self.vocab_list) - self.assertEqual(bst_result, self.greedy_result[1]) - - def test_beam_search_decoder_1(self): - beam_result = decoder.ctc_beam_search_decoder( - probs_seq=self.probs_seq1, - beam_size=self.beam_size, - vocabulary=self.vocab_list) - self.assertEqual(beam_result[0][1], self.beam_search_result[0]) - - def test_beam_search_decoder_2(self): - beam_result = decoder.ctc_beam_search_decoder( - probs_seq=self.probs_seq2, - beam_size=self.beam_size, - vocabulary=self.vocab_list) - self.assertEqual(beam_result[0][1], self.beam_search_result[1]) - - def test_beam_search_decoder_batch(self): - beam_results = decoder.ctc_beam_search_decoder_batch( - probs_split=[self.probs_seq1, self.probs_seq2], - beam_size=self.beam_size, - vocabulary=self.vocab_list, - num_processes=24) - self.assertEqual(beam_results[0][0][1], self.beam_search_result[0]) - self.assertEqual(beam_results[1][0][1], self.beam_search_result[1]) - - -if __name__ == '__main__': - unittest.main() diff --git a/deep_speech_2/deploy/_init_paths.py b/deep_speech_2/deploy/_init_paths.py deleted file mode 100644 index ddabb535be682d95c3c8b73003ea30eed06ca0b0..0000000000000000000000000000000000000000 --- a/deep_speech_2/deploy/_init_paths.py +++ /dev/null @@ -1,19 +0,0 @@ -"""Set up paths for DS2""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os.path -import sys - - -def add_path(path): - if path not in sys.path: - sys.path.insert(0, path) - - -this_dir = os.path.dirname(__file__) - -# Add project path to PYTHONPATH -proj_path = os.path.join(this_dir, '..') -add_path(proj_path) diff --git a/deep_speech_2/deploy/demo_client.py b/deep_speech_2/deploy/demo_client.py deleted file mode 100644 index ddf4dd1bf3f5ea62661e181e0dd2fb3f3b1379c6..0000000000000000000000000000000000000000 --- a/deep_speech_2/deploy/demo_client.py +++ /dev/null @@ -1,94 +0,0 @@ -"""Client-end for the ASR demo.""" -from pynput import keyboard -import struct -import socket -import sys -import argparse -import pyaudio - -parser = argparse.ArgumentParser(description=__doc__) -parser.add_argument( - "--host_ip", - default="localhost", - type=str, - help="Server IP address. (default: %(default)s)") -parser.add_argument( - "--host_port", - default=8086, - type=int, - help="Server Port. (default: %(default)s)") -args = parser.parse_args() - -is_recording = False -enable_trigger_record = True - - -def on_press(key): - """On-press keyboard callback function.""" - global is_recording, enable_trigger_record - if key == keyboard.Key.space: - if (not is_recording) and enable_trigger_record: - sys.stdout.write("Start Recording ... ") - sys.stdout.flush() - is_recording = True - - -def on_release(key): - """On-release keyboard callback function.""" - global is_recording, enable_trigger_record - if key == keyboard.Key.esc: - return False - elif key == keyboard.Key.space: - if is_recording == True: - is_recording = False - - -data_list = [] - - -def callback(in_data, frame_count, time_info, status): - """Audio recorder's stream callback function.""" - global data_list, is_recording, enable_trigger_record - if is_recording: - data_list.append(in_data) - enable_trigger_record = False - elif len(data_list) > 0: - # Connect to server and send data - sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM) - sock.connect((args.host_ip, args.host_port)) - sent = ''.join(data_list) - sock.sendall(struct.pack('>i', len(sent)) + sent) - print('Speech[length=%d] Sent.' % len(sent)) - # Receive data from the server and shut down - received = sock.recv(1024) - print "Recognition Results: {}".format(received) - sock.close() - data_list = [] - enable_trigger_record = True - return (in_data, pyaudio.paContinue) - - -def main(): - # prepare audio recorder - p = pyaudio.PyAudio() - stream = p.open( - format=pyaudio.paInt32, - channels=1, - rate=16000, - input=True, - stream_callback=callback) - stream.start_stream() - - # prepare keyboard listener - with keyboard.Listener( - on_press=on_press, on_release=on_release) as listener: - listener.join() - - # close up - stream.stop_stream() - stream.close() - p.terminate() - - -if __name__ == "__main__": - main() diff --git a/deep_speech_2/deploy/demo_server.py b/deep_speech_2/deploy/demo_server.py deleted file mode 100644 index 3e81c0c5b92a873cc5854c727429c75a3ae1fb0d..0000000000000000000000000000000000000000 --- a/deep_speech_2/deploy/demo_server.py +++ /dev/null @@ -1,205 +0,0 @@ -"""Server-end for the ASR demo.""" -import os -import time -import random -import argparse -import functools -from time import gmtime, strftime -import SocketServer -import struct -import wave -import paddle.v2 as paddle -import _init_paths -from data_utils.data import DataGenerator -from model_utils.model import DeepSpeech2Model -from data_utils.utility import read_manifest -from utils.utility import add_arguments, print_arguments - -parser = argparse.ArgumentParser(description=__doc__) -add_arg = functools.partial(add_arguments, argparser=parser) -# yapf: disable -add_arg('host_port', int, 8086, "Server's IP port.") -add_arg('beam_size', int, 500, "Beam search width.") -add_arg('num_conv_layers', int, 2, "# of convolution layers.") -add_arg('num_rnn_layers', int, 3, "# of recurrent layers.") -add_arg('rnn_layer_size', int, 2048, "# of recurrent cells per layer.") -add_arg('alpha', float, 2.15, "Coef of LM for beam search.") -add_arg('beta', float, 0.35, "Coef of WC for beam search.") -add_arg('cutoff_prob', float, 1.0, "Cutoff probability for pruning.") -add_arg('cutoff_top_n', int, 40, "Cutoff number for pruning.") -add_arg('use_gru', bool, False, "Use GRUs instead of simple RNNs.") -add_arg('use_gpu', bool, True, "Use GPU or not.") -add_arg('share_rnn_weights',bool, True, "Share input-hidden weights across " - "bi-directional RNNs. Not for GRU.") -add_arg('host_ip', str, - 'localhost', - "Server's IP address.") -add_arg('speech_save_dir', str, - 'demo_cache', - "Directory to save demo audios.") -add_arg('warmup_manifest', str, - 'data/librispeech/manifest.test-clean', - "Filepath of manifest to warm up.") -add_arg('mean_std_path', str, - 'data/librispeech/mean_std.npz', - "Filepath of normalizer's mean & std.") -add_arg('vocab_path', str, - 'data/librispeech/eng_vocab.txt', - "Filepath of vocabulary.") -add_arg('model_path', str, - './checkpoints/libri/params.latest.tar.gz', - "If None, the training starts from scratch, " - "otherwise, it resumes from the pre-trained model.") -add_arg('lang_model_path', str, - 'lm/data/common_crawl_00.prune01111.trie.klm', - "Filepath for language model.") -add_arg('decoding_method', str, - 'ctc_beam_search', - "Decoding method. Options: ctc_beam_search, ctc_greedy", - choices = ['ctc_beam_search', 'ctc_greedy']) -add_arg('specgram_type', str, - 'linear', - "Audio feature type. Options: linear, mfcc.", - choices=['linear', 'mfcc']) -# yapf: disable -args = parser.parse_args() - - -class AsrTCPServer(SocketServer.TCPServer): - """The ASR TCP Server.""" - - def __init__(self, - server_address, - RequestHandlerClass, - speech_save_dir, - audio_process_handler, - bind_and_activate=True): - self.speech_save_dir = speech_save_dir - self.audio_process_handler = audio_process_handler - SocketServer.TCPServer.__init__( - self, server_address, RequestHandlerClass, bind_and_activate=True) - - -class AsrRequestHandler(SocketServer.BaseRequestHandler): - """The ASR request handler.""" - - def handle(self): - # receive data through TCP socket - chunk = self.request.recv(1024) - target_len = struct.unpack('>i', chunk[:4])[0] - data = chunk[4:] - while len(data) < target_len: - chunk = self.request.recv(1024) - data += chunk - # write to file - filename = self._write_to_file(data) - - print("Received utterance[length=%d] from %s, saved to %s." % - (len(data), self.client_address[0], filename)) - start_time = time.time() - transcript = self.server.audio_process_handler(filename) - finish_time = time.time() - print("Response Time: %f, Transcript: %s" % - (finish_time - start_time, transcript)) - self.request.sendall(transcript.encode('utf-8')) - - def _write_to_file(self, data): - # prepare save dir and filename - if not os.path.exists(self.server.speech_save_dir): - os.mkdir(self.server.speech_save_dir) - timestamp = strftime("%Y%m%d%H%M%S", gmtime()) - out_filename = os.path.join( - self.server.speech_save_dir, - timestamp + "_" + self.client_address[0] + ".wav") - # write to wav file - file = wave.open(out_filename, 'wb') - file.setnchannels(1) - file.setsampwidth(4) - file.setframerate(16000) - file.writeframes(data) - file.close() - return out_filename - - -def warm_up_test(audio_process_handler, - manifest_path, - num_test_cases, - random_seed=0): - """Warming-up test.""" - manifest = read_manifest(manifest_path) - rng = random.Random(random_seed) - samples = rng.sample(manifest, num_test_cases) - for idx, sample in enumerate(samples): - print("Warm-up Test Case %d: %s", idx, sample['audio_filepath']) - start_time = time.time() - transcript = audio_process_handler(sample['audio_filepath']) - finish_time = time.time() - print("Response Time: %f, Transcript: %s" % - (finish_time - start_time, transcript)) - - -def start_server(): - """Start the ASR server""" - # prepare data generator - data_generator = DataGenerator( - vocab_filepath=args.vocab_path, - mean_std_filepath=args.mean_std_path, - augmentation_config='{}', - specgram_type=args.specgram_type, - num_threads=1, - keep_transcription_text=True) - # prepare ASR model - ds2_model = DeepSpeech2Model( - vocab_size=data_generator.vocab_size, - num_conv_layers=args.num_conv_layers, - num_rnn_layers=args.num_rnn_layers, - rnn_layer_size=args.rnn_layer_size, - use_gru=args.use_gru, - pretrained_model_path=args.model_path, - share_rnn_weights=args.share_rnn_weights) - - vocab_list = [chars.encode("utf-8") for chars in data_generator.vocab_list] - - # prepare ASR inference handler - def file_to_transcript(filename): - feature = data_generator.process_utterance(filename, "") - result_transcript = ds2_model.infer_batch( - infer_data=[feature], - decoding_method=args.decoding_method, - beam_alpha=args.alpha, - beam_beta=args.beta, - beam_size=args.beam_size, - cutoff_prob=args.cutoff_prob, - cutoff_top_n=args.cutoff_top_n, - vocab_list=vocab_list, - language_model_path=args.lang_model_path, - num_processes=1) - return result_transcript[0] - - # warming up with utterrances sampled from Librispeech - print('-----------------------------------------------------------') - print('Warming up ...') - warm_up_test( - audio_process_handler=file_to_transcript, - manifest_path=args.warmup_manifest, - num_test_cases=3) - print('-----------------------------------------------------------') - - # start the server - server = AsrTCPServer( - server_address=(args.host_ip, args.host_port), - RequestHandlerClass=AsrRequestHandler, - speech_save_dir=args.speech_save_dir, - audio_process_handler=file_to_transcript) - print("ASR Server Started.") - server.serve_forever() - - -def main(): - print_arguments(args) - paddle.init(use_gpu=args.use_gpu, trainer_count=1) - start_server() - - -if __name__ == "__main__": - main() diff --git a/deep_speech_2/docs/images/multi_gpu_speedup.png b/deep_speech_2/docs/images/multi_gpu_speedup.png deleted file mode 100755 index 57a803bac8a6c793548abf71db2899aeac829fa1..0000000000000000000000000000000000000000 Binary files a/deep_speech_2/docs/images/multi_gpu_speedup.png and /dev/null differ diff --git a/deep_speech_2/docs/images/tuning_error_surface.png b/deep_speech_2/docs/images/tuning_error_surface.png deleted file mode 100644 index 2204cee2f5204d1d2d2e53fab8cdd0a1cb9ac47d..0000000000000000000000000000000000000000 Binary files a/deep_speech_2/docs/images/tuning_error_surface.png and /dev/null differ diff --git a/deep_speech_2/examples/aishell/run_data.sh b/deep_speech_2/examples/aishell/run_data.sh deleted file mode 100644 index eb0388d84f6429b02b0c58108c2543390d89c440..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/aishell/run_data.sh +++ /dev/null @@ -1,42 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download data, generate manifests -PYTHONPATH=.:$PYTHONPATH python data/aishell/aishell.py \ ---manifest_prefix='data/aishell/manifest' \ ---target_dir='~/.cache/paddle/dataset/speech/Aishell' - -if [ $? -ne 0 ]; then - echo "Prepare Aishell failed. Terminated." - exit 1 -fi - - -# build vocabulary -python tools/build_vocab.py \ ---count_threshold=0 \ ---vocab_path='data/aishell/vocab.txt' \ ---manifest_paths 'data/aishell/manifest.train' 'data/aishell/manifest.dev' - -if [ $? -ne 0 ]; then - echo "Build vocabulary failed. Terminated." - exit 1 -fi - - -# compute mean and stddev for normalizer -python tools/compute_mean_std.py \ ---manifest_path='data/aishell/manifest.train' \ ---num_samples=2000 \ ---specgram_type='linear' \ ---output_path='data/aishell/mean_std.npz' - -if [ $? -ne 0 ]; then - echo "Compute mean and stddev failed. Terminated." - exit 1 -fi - - -echo "Aishell data preparation done." -exit 0 diff --git a/deep_speech_2/examples/aishell/run_infer.sh b/deep_speech_2/examples/aishell/run_infer.sh deleted file mode 100644 index 404555e8b10ef7befff3b67a94b1cadad858e592..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/aishell/run_infer.sh +++ /dev/null @@ -1,46 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_ch.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# infer -CUDA_VISIBLE_DEVICES=0 \ -python -u infer.py \ ---num_samples=10 \ ---trainer_count=1 \ ---beam_size=300 \ ---num_proc_bsearch=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=1024 \ ---alpha=1.4 \ ---beta=2.4 \ ---cutoff_prob=0.99 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=False \ ---infer_manifest='data/aishell/manifest.test' \ ---mean_std_path='data/aishell/mean_std.npz' \ ---vocab_path='data/aishell/vocab.txt' \ ---model_path='checkpoints/aishell/params.latest.tar.gz' \ ---lang_model_path='models/lm/zh_giga.no_cna_cmn.prune01244.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='cer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in inference!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/aishell/run_infer_golden.sh b/deep_speech_2/examples/aishell/run_infer_golden.sh deleted file mode 100644 index 4701bdaaceae5a345610451ad6b0197dec530617..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/aishell/run_infer_golden.sh +++ /dev/null @@ -1,55 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_ch.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# download well-trained model -cd models/aishell > /dev/null -sh download_model.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# infer -CUDA_VISIBLE_DEVICES=0 \ -python -u infer.py \ ---num_samples=10 \ ---trainer_count=1 \ ---beam_size=300 \ ---num_proc_bsearch=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=1024 \ ---alpha=1.4 \ ---beta=2.4 \ ---cutoff_prob=0.99 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=False \ ---infer_manifest='data/aishell/manifest.test' \ ---mean_std_path='models/aishell/mean_std.npz' \ ---vocab_path='models/aishell/vocab.txt' \ ---model_path='models/aishell/params.tar.gz' \ ---lang_model_path='models/lm/zh_giga.no_cna_cmn.prune01244.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='cer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in inference!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/aishell/run_test.sh b/deep_speech_2/examples/aishell/run_test.sh deleted file mode 100644 index feec95cb45f1b6b456d0dfad1eabc97e9e59045f..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/aishell/run_test.sh +++ /dev/null @@ -1,47 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_ch.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# evaluate model -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u test.py \ ---batch_size=128 \ ---trainer_count=8 \ ---beam_size=300 \ ---num_proc_bsearch=8 \ ---num_proc_data=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=1024 \ ---alpha=1.4 \ ---beta=2.4 \ ---cutoff_prob=0.99 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=False \ ---test_manifest='data/aishell/manifest.test' \ ---mean_std_path='data/aishell/mean_std.npz' \ ---vocab_path='data/aishell/vocab.txt' \ ---model_path='checkpoints/aishell/params.latest.tar.gz' \ ---lang_model_path='models/lm/zh_giga.no_cna_cmn.prune01244.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='cer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in evaluation!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/aishell/run_test_golden.sh b/deep_speech_2/examples/aishell/run_test_golden.sh deleted file mode 100644 index 387d54f395449e4ea9b3bdc893fcc765ff0f4d2c..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/aishell/run_test_golden.sh +++ /dev/null @@ -1,56 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_ch.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# download well-trained model -cd models/aishell > /dev/null -sh download_model.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# evaluate model -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u test.py \ ---batch_size=128 \ ---trainer_count=8 \ ---beam_size=300 \ ---num_proc_bsearch=8 \ ---num_proc_data=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=1024 \ ---alpha=1.4 \ ---beta=2.4 \ ---cutoff_prob=0.99 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=False \ ---test_manifest='data/aishell/manifest.test' \ ---mean_std_path='models/aishell/mean_std.npz' \ ---vocab_path='models/aishell/vocab.txt' \ ---model_path='models/aishell/params.tar.gz' \ ---lang_model_path='models/lm/zh_giga.no_cna_cmn.prune01244.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='cer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in evaluation!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/aishell/run_train.sh b/deep_speech_2/examples/aishell/run_train.sh deleted file mode 100644 index 077fabcd8d21b235058806dbb54445c48f20989b..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/aishell/run_train.sh +++ /dev/null @@ -1,41 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# train model -# if you wish to resume from an exists model, uncomment --init_model_path -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u train.py \ ---batch_size=64 \ ---trainer_count=8 \ ---num_passes=50 \ ---num_proc_data=16 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=1024 \ ---num_iter_print=100 \ ---learning_rate=5e-4 \ ---max_duration=27.0 \ ---min_duration=0.0 \ ---test_off=False \ ---use_sortagrad=True \ ---use_gru=False \ ---use_gpu=True \ ---is_local=True \ ---share_rnn_weights=False \ ---train_manifest='data/aishell/manifest.train' \ ---dev_manifest='data/aishell/manifest.dev' \ ---mean_std_path='data/aishell/mean_std.npz' \ ---vocab_path='data/aishell/vocab.txt' \ ---output_model_dir='./checkpoints/aishell' \ ---augment_conf_path='conf/augmentation.config' \ ---specgram_type='linear' \ ---shuffle_method='batch_shuffle_clipped' - -if [ $? -ne 0 ]; then - echo "Failed in training!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/deploy_demo/run_demo_client.sh b/deep_speech_2/examples/deploy_demo/run_demo_client.sh deleted file mode 100644 index 6ae4ddcab92e8cf3b207987513b438cf73e194ee..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/deploy_demo/run_demo_client.sh +++ /dev/null @@ -1,17 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# start demo client -CUDA_VISIBLE_DEVICES=0 \ -python -u deploy/demo_client.py \ ---host_ip='localhost' \ ---host_port=8086 \ - -if [ $? -ne 0 ]; then - echo "Failed in starting demo client!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/deploy_demo/run_english_demo_server.sh b/deep_speech_2/examples/deploy_demo/run_english_demo_server.sh deleted file mode 100644 index 67532770c137d33b75a35f1646f85d6fe4f3eb5e..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/deploy_demo/run_english_demo_server.sh +++ /dev/null @@ -1,54 +0,0 @@ -#! /usr/bin/env bash -# TODO: replace the model with a mandarin model - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# download well-trained model -cd models/baidu_en8k > /dev/null -sh download_model.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# start demo server -CUDA_VISIBLE_DEVICES=0 \ -python -u deploy/demo_server.py \ ---host_ip='localhost' \ ---host_port=8086 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=1024 \ ---alpha=1.15 \ ---beta=0.15 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=True \ ---use_gpu=True \ ---share_rnn_weights=False \ ---speech_save_dir='demo_cache' \ ---warmup_manifest='data/tiny/manifest.test-clean' \ ---mean_std_path='models/baidu_en8k/mean_std.npz' \ ---vocab_path='models/baidu_en8k/vocab.txt' \ ---model_path='models/baidu_en8k/params.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in starting demo server!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/librispeech/run_data.sh b/deep_speech_2/examples/librispeech/run_data.sh deleted file mode 100644 index 12f2dc6d553695c8a61ac9600c3761ba84c4475c..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/librispeech/run_data.sh +++ /dev/null @@ -1,45 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download data, generate manifests -PYTHONPATH=.:$PYTHONPATH python data/librispeech/librispeech.py \ ---manifest_prefix='data/librispeech/manifest' \ ---target_dir='~/.cache/paddle/dataset/speech/Libri' \ ---full_download='True' - -if [ $? -ne 0 ]; then - echo "Prepare LibriSpeech failed. Terminated." - exit 1 -fi - -cat data/librispeech/manifest.train-* | shuf > data/librispeech/manifest.train - - -# build vocabulary -python tools/build_vocab.py \ ---count_threshold=0 \ ---vocab_path='data/librispeech/vocab.txt' \ ---manifest_paths='data/librispeech/manifest.train' - -if [ $? -ne 0 ]; then - echo "Build vocabulary failed. Terminated." - exit 1 -fi - - -# compute mean and stddev for normalizer -python tools/compute_mean_std.py \ ---manifest_path='data/librispeech/manifest.train' \ ---num_samples=2000 \ ---specgram_type='linear' \ ---output_path='data/librispeech/mean_std.npz' - -if [ $? -ne 0 ]; then - echo "Compute mean and stddev failed. Terminated." - exit 1 -fi - - -echo "LibriSpeech Data preparation done." -exit 0 diff --git a/deep_speech_2/examples/librispeech/run_infer.sh b/deep_speech_2/examples/librispeech/run_infer.sh deleted file mode 100644 index 2df5b6cc410bbb7a3722a568378aec8be2688c64..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/librispeech/run_infer.sh +++ /dev/null @@ -1,46 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# infer -CUDA_VISIBLE_DEVICES=0 \ -python -u infer.py \ ---num_samples=10 \ ---trainer_count=1 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---infer_manifest='data/librispeech/manifest.test-clean' \ ---mean_std_path='data/librispeech/mean_std.npz' \ ---vocab_path='data/librispeech/vocab.txt' \ ---model_path='checkpoints/libri/params.latest.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in inference!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/librispeech/run_infer_golden.sh b/deep_speech_2/examples/librispeech/run_infer_golden.sh deleted file mode 100644 index c407cabe44ef906884f8447692dff8458ebe8ffc..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/librispeech/run_infer_golden.sh +++ /dev/null @@ -1,55 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# download well-trained model -cd models/librispeech > /dev/null -sh download_model.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# infer -CUDA_VISIBLE_DEVICES=0 \ -python -u infer.py \ ---num_samples=10 \ ---trainer_count=1 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---infer_manifest='data/librispeech/manifest.test-clean' \ ---mean_std_path='models/librispeech/mean_std.npz' \ ---vocab_path='models/librispeech/vocab.txt' \ ---model_path='models/librispeech/params.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in inference!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/librispeech/run_test.sh b/deep_speech_2/examples/librispeech/run_test.sh deleted file mode 100644 index 0a76704d75d02c2a572da1d1504aed0d4289f964..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/librispeech/run_test.sh +++ /dev/null @@ -1,47 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# evaluate model -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u test.py \ ---batch_size=128 \ ---trainer_count=8 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_proc_data=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---test_manifest='data/librispeech/manifest.test-clean' \ ---mean_std_path='data/librispeech/mean_std.npz' \ ---vocab_path='data/librispeech/vocab.txt' \ ---model_path='checkpoints/libri/params.latest.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in evaluation!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/librispeech/run_test_golden.sh b/deep_speech_2/examples/librispeech/run_test_golden.sh deleted file mode 100644 index 3e7e3b4c5c4362419c4e2cb4a6462d0ff511be46..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/librispeech/run_test_golden.sh +++ /dev/null @@ -1,56 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# download well-trained model -cd models/librispeech > /dev/null -sh download_model.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# evaluate model -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u test.py \ ---batch_size=128 \ ---trainer_count=8 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_proc_data=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---test_manifest='data/librispeech/manifest.test-clean' \ ---mean_std_path='models/librispeech/mean_std.npz' \ ---vocab_path='models/librispeech/vocab.txt' \ ---model_path='models/librispeech/params.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in evaluation!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/librispeech/run_train.sh b/deep_speech_2/examples/librispeech/run_train.sh deleted file mode 100644 index 87e08721b3306de3c6a91d2c11993deaf09da967..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/librispeech/run_train.sh +++ /dev/null @@ -1,41 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# train model -# if you wish to resume from an exists model, uncomment --init_model_path -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u train.py \ ---batch_size=160 \ ---trainer_count=8 \ ---num_passes=50 \ ---num_proc_data=16 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---num_iter_print=100 \ ---learning_rate=5e-4 \ ---max_duration=27.0 \ ---min_duration=0.0 \ ---test_off=False \ ---use_sortagrad=True \ ---use_gru=False \ ---use_gpu=True \ ---is_local=True \ ---share_rnn_weights=True \ ---train_manifest='data/librispeech/manifest.train' \ ---dev_manifest='data/librispeech/manifest.dev-clean' \ ---mean_std_path='data/librispeech/mean_std.npz' \ ---vocab_path='data/librispeech/vocab.txt' \ ---output_model_dir='./checkpoints/libri' \ ---augment_conf_path='conf/augmentation.config' \ ---specgram_type='linear' \ ---shuffle_method='batch_shuffle_clipped' - -if [ $? -ne 0 ]; then - echo "Failed in training!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/librispeech/run_tune.sh b/deep_speech_2/examples/librispeech/run_tune.sh deleted file mode 100644 index c3695d1cb2ccb3bee206bc6cd432e49eba26e3bf..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/librispeech/run_tune.sh +++ /dev/null @@ -1,41 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# grid-search for hyper-parameters in language model -CUDA_VISIBLE_DEVICES=0,1,2,3 \ -python -u tools/tune.py \ ---num_batches=-1 \ ---batch_size=128 \ ---trainer_count=8 \ ---beam_size=500 \ ---num_proc_bsearch=12 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---num_alphas=45 \ ---num_betas=8 \ ---alpha_from=1.0 \ ---alpha_to=3.2 \ ---beta_from=0.1 \ ---beta_to=0.45 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---tune_manifest='data/librispeech/manifest.dev-clean' \ ---mean_std_path='data/librispeech/mean_std.npz' \ ---vocab_path='models/librispeech/vocab.txt' \ ---model_path='models/librispeech/params.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in tuning!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/tiny/run_data.sh b/deep_speech_2/examples/tiny/run_data.sh deleted file mode 100644 index ba55d284a8fd583bd60d0fee3e6f0ad40008ff9a..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/tiny/run_data.sh +++ /dev/null @@ -1,51 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# prepare folder -if [ ! -e data/tiny ]; then - mkdir data/tiny -fi - - -# download data, generate manifests -PYTHONPATH=.:$PYTHONPATH python data/librispeech/librispeech.py \ ---manifest_prefix='data/tiny/manifest' \ ---target_dir='~/.cache/paddle/dataset/speech/libri' \ ---full_download='False' - -if [ $? -ne 0 ]; then - echo "Prepare LibriSpeech failed. Terminated." - exit 1 -fi - -head -n 64 data/tiny/manifest.dev-clean > data/tiny/manifest.tiny - - -# build vocabulary -python tools/build_vocab.py \ ---count_threshold=0 \ ---vocab_path='data/tiny/vocab.txt' \ ---manifest_paths='data/tiny/manifest.dev-clean' - -if [ $? -ne 0 ]; then - echo "Build vocabulary failed. Terminated." - exit 1 -fi - - -# compute mean and stddev for normalizer -python tools/compute_mean_std.py \ ---manifest_path='data/tiny/manifest.tiny' \ ---num_samples=64 \ ---specgram_type='linear' \ ---output_path='data/tiny/mean_std.npz' - -if [ $? -ne 0 ]; then - echo "Compute mean and stddev failed. Terminated." - exit 1 -fi - - -echo "Tiny data preparation done." -exit 0 diff --git a/deep_speech_2/examples/tiny/run_infer.sh b/deep_speech_2/examples/tiny/run_infer.sh deleted file mode 100644 index 3a345f2fff98512203828a7d5fbcd27fbb84e69e..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/tiny/run_infer.sh +++ /dev/null @@ -1,46 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# infer -CUDA_VISIBLE_DEVICES=0 \ -python -u infer.py \ ---num_samples=10 \ ---trainer_count=1 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---infer_manifest='data/tiny/manifest.tiny' \ ---mean_std_path='data/tiny/mean_std.npz' \ ---vocab_path='data/tiny/vocab.txt' \ ---model_path='checkpoints/tiny/params.pass-19.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in inference!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/tiny/run_infer_golden.sh b/deep_speech_2/examples/tiny/run_infer_golden.sh deleted file mode 100644 index 72a8be064973a4a6dd52a8a0575759f06720df33..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/tiny/run_infer_golden.sh +++ /dev/null @@ -1,55 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# download well-trained model -cd models/librispeech > /dev/null -sh download_model.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# infer -CUDA_VISIBLE_DEVICES=0 \ -python -u infer.py \ ---num_samples=10 \ ---trainer_count=1 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---infer_manifest='data/tiny/manifest.test-clean' \ ---mean_std_path='models/librispeech/mean_std.npz' \ ---vocab_path='models/librispeech/vocab.txt' \ ---model_path='models/librispeech/params.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in inference!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/tiny/run_test.sh b/deep_speech_2/examples/tiny/run_test.sh deleted file mode 100644 index a58f5d100d97a8067089bf55149c0235c20e3a96..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/tiny/run_test.sh +++ /dev/null @@ -1,47 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# evaluate model -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u test.py \ ---batch_size=16 \ ---trainer_count=8 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_proc_data=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---test_manifest='data/tiny/manifest.tiny' \ ---mean_std_path='data/tiny/mean_std.npz' \ ---vocab_path='data/tiny/vocab.txt' \ ---model_path='checkpoints/params.pass-19.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in evaluation!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/tiny/run_test_golden.sh b/deep_speech_2/examples/tiny/run_test_golden.sh deleted file mode 100644 index 8d3d25c5cea5f1963830fad65218cc6c5d0fcfe0..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/tiny/run_test_golden.sh +++ /dev/null @@ -1,56 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# download language model -cd models/lm > /dev/null -sh download_lm_en.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# download well-trained model -cd models/librispeech > /dev/null -sh download_model.sh -if [ $? -ne 0 ]; then - exit 1 -fi -cd - > /dev/null - - -# evaluate model -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u test.py \ ---batch_size=128 \ ---trainer_count=8 \ ---beam_size=500 \ ---num_proc_bsearch=8 \ ---num_proc_data=8 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---alpha=2.15 \ ---beta=0.35 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---test_manifest='data/tiny/manifest.test-clean' \ ---mean_std_path='models/librispeech/mean_std.npz' \ ---vocab_path='models/librispeech/vocab.txt' \ ---model_path='models/librispeech/params.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---decoding_method='ctc_beam_search' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in evaluation!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/tiny/run_train.sh b/deep_speech_2/examples/tiny/run_train.sh deleted file mode 100644 index e03a8aff04965692ae8c69e574b1cd94ea4677fc..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/tiny/run_train.sh +++ /dev/null @@ -1,41 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# train model -# if you wish to resume from an exists model, uncomment --init_model_path -CUDA_VISIBLE_DEVICES=0,1,2,3 \ -python -u train.py \ ---batch_size=16 \ ---trainer_count=4 \ ---num_passes=20 \ ---num_proc_data=1 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---num_iter_print=100 \ ---learning_rate=1e-5 \ ---max_duration=27.0 \ ---min_duration=0.0 \ ---test_off=False \ ---use_sortagrad=True \ ---use_gru=False \ ---use_gpu=True \ ---is_local=True \ ---share_rnn_weights=True \ ---train_manifest='data/tiny/manifest.tiny' \ ---dev_manifest='data/tiny/manifest.tiny' \ ---mean_std_path='data/tiny/mean_std.npz' \ ---vocab_path='data/tiny/vocab.txt' \ ---output_model_dir='./checkpoints/tiny' \ ---augment_conf_path='conf/augmentation.config' \ ---specgram_type='linear' \ ---shuffle_method='batch_shuffle_clipped' - -if [ $? -ne 0 ]; then - echo "Fail in training!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/examples/tiny/run_tune.sh b/deep_speech_2/examples/tiny/run_tune.sh deleted file mode 100644 index 89f8adf45312517f3b118309eade15bda617125a..0000000000000000000000000000000000000000 --- a/deep_speech_2/examples/tiny/run_tune.sh +++ /dev/null @@ -1,41 +0,0 @@ -#! /usr/bin/env bash - -cd ../.. > /dev/null - -# grid-search for hyper-parameters in language model -CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \ -python -u tools/tune.py \ ---num_batches=1 \ ---batch_size=24 \ ---trainer_count=8 \ ---beam_size=500 \ ---num_proc_bsearch=12 \ ---num_conv_layers=2 \ ---num_rnn_layers=3 \ ---rnn_layer_size=2048 \ ---num_alphas=45 \ ---num_betas=8 \ ---alpha_from=1.0 \ ---alpha_to=3.2 \ ---beta_from=0.1 \ ---beta_to=0.45 \ ---cutoff_prob=1.0 \ ---cutoff_top_n=40 \ ---use_gru=False \ ---use_gpu=True \ ---share_rnn_weights=True \ ---tune_manifest='data/tiny/manifest.tiny' \ ---mean_std_path='data/tiny/mean_std.npz' \ ---vocab_path='data/tiny/vocab.txt' \ ---model_path='checkpoints/params.pass-9.tar.gz' \ ---lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \ ---error_rate_type='wer' \ ---specgram_type='linear' - -if [ $? -ne 0 ]; then - echo "Failed in tuning!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/infer.py b/deep_speech_2/infer.py deleted file mode 100644 index 32d15f1265f4a067588522bd396f52d5b8edf423..0000000000000000000000000000000000000000 --- a/deep_speech_2/infer.py +++ /dev/null @@ -1,126 +0,0 @@ -"""Inferer for DeepSpeech2 model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import argparse -import functools -import paddle.v2 as paddle -from data_utils.data import DataGenerator -from model_utils.model import DeepSpeech2Model -from utils.error_rate import wer, cer -from utils.utility import add_arguments, print_arguments - -parser = argparse.ArgumentParser(description=__doc__) -add_arg = functools.partial(add_arguments, argparser=parser) -# yapf: disable -add_arg('num_samples', int, 10, "# of samples to infer.") -add_arg('trainer_count', int, 8, "# of Trainers (CPUs or GPUs).") -add_arg('beam_size', int, 500, "Beam search width.") -add_arg('num_proc_bsearch', int, 8, "# of CPUs for beam search.") -add_arg('num_conv_layers', int, 2, "# of convolution layers.") -add_arg('num_rnn_layers', int, 3, "# of recurrent layers.") -add_arg('rnn_layer_size', int, 2048, "# of recurrent cells per layer.") -add_arg('alpha', float, 2.15, "Coef of LM for beam search.") -add_arg('beta', float, 0.35, "Coef of WC for beam search.") -add_arg('cutoff_prob', float, 1.0, "Cutoff probability for pruning.") -add_arg('cutoff_top_n', int, 40, "Cutoff number for pruning.") -add_arg('use_gru', bool, False, "Use GRUs instead of simple RNNs.") -add_arg('use_gpu', bool, True, "Use GPU or not.") -add_arg('share_rnn_weights',bool, True, "Share input-hidden weights across " - "bi-directional RNNs. Not for GRU.") -add_arg('infer_manifest', str, - 'data/librispeech/manifest.dev-clean', - "Filepath of manifest to infer.") -add_arg('mean_std_path', str, - 'data/librispeech/mean_std.npz', - "Filepath of normalizer's mean & std.") -add_arg('vocab_path', str, - 'data/librispeech/vocab.txt', - "Filepath of vocabulary.") -add_arg('lang_model_path', str, - 'models/lm/common_crawl_00.prune01111.trie.klm', - "Filepath for language model.") -add_arg('model_path', str, - './checkpoints/libri/params.latest.tar.gz', - "If None, the training starts from scratch, " - "otherwise, it resumes from the pre-trained model.") -add_arg('decoding_method', str, - 'ctc_beam_search', - "Decoding method. Options: ctc_beam_search, ctc_greedy", - choices = ['ctc_beam_search', 'ctc_greedy']) -add_arg('error_rate_type', str, - 'wer', - "Error rate type for evaluation.", - choices=['wer', 'cer']) -add_arg('specgram_type', str, - 'linear', - "Audio feature type. Options: linear, mfcc.", - choices=['linear', 'mfcc']) -# yapf: disable -args = parser.parse_args() - - -def infer(): - """Inference for DeepSpeech2.""" - data_generator = DataGenerator( - vocab_filepath=args.vocab_path, - mean_std_filepath=args.mean_std_path, - augmentation_config='{}', - specgram_type=args.specgram_type, - num_threads=1, - keep_transcription_text=True, - num_conv_layers=args.num_conv_layers) - batch_reader = data_generator.batch_reader_creator( - manifest_path=args.infer_manifest, - batch_size=args.num_samples, - min_batch_size=1, - sortagrad=False, - shuffle_method=None) - infer_data = batch_reader().next() - - ds2_model = DeepSpeech2Model( - vocab_size=data_generator.vocab_size, - num_conv_layers=args.num_conv_layers, - num_rnn_layers=args.num_rnn_layers, - rnn_layer_size=args.rnn_layer_size, - use_gru=args.use_gru, - pretrained_model_path=args.model_path, - share_rnn_weights=args.share_rnn_weights) - - # decoders only accept string encoded in utf-8 - vocab_list = [chars.encode("utf-8") for chars in data_generator.vocab_list] - - result_transcripts = ds2_model.infer_batch( - infer_data=infer_data, - decoding_method=args.decoding_method, - beam_alpha=args.alpha, - beam_beta=args.beta, - beam_size=args.beam_size, - cutoff_prob=args.cutoff_prob, - cutoff_top_n=args.cutoff_top_n, - vocab_list=vocab_list, - language_model_path=args.lang_model_path, - num_processes=args.num_proc_bsearch, - feeding_dict=data_generator.feeding) - - error_rate_func = cer if args.error_rate_type == 'cer' else wer - target_transcripts = [data[1] for data in infer_data] - for target, result in zip(target_transcripts, result_transcripts): - print("\nTarget Transcription: %s\nOutput Transcription: %s" % - (target, result)) - print("Current error rate [%s] = %f" % - (args.error_rate_type, error_rate_func(target, result))) - - ds2_model.logger.info("finish inference") - -def main(): - print_arguments(args) - paddle.init(use_gpu=args.use_gpu, - rnn_use_batch=True, - trainer_count=args.trainer_count) - infer() - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/model_utils/__init__.py b/deep_speech_2/model_utils/__init__.py deleted file mode 100644 index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0000000000000000000000000000000000000000 diff --git a/deep_speech_2/model_utils/model.py b/deep_speech_2/model_utils/model.py deleted file mode 100644 index 26aa1470aa61f8e5c13edaba88b8acb8bf619f7b..0000000000000000000000000000000000000000 --- a/deep_speech_2/model_utils/model.py +++ /dev/null @@ -1,307 +0,0 @@ -"""Contains DeepSpeech2 model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import sys -import os -import time -import logging -import gzip -from distutils.dir_util import mkpath -import paddle.v2 as paddle -from decoders.swig_wrapper import Scorer -from decoders.swig_wrapper import ctc_greedy_decoder -from decoders.swig_wrapper import ctc_beam_search_decoder_batch -from model_utils.network import deep_speech_v2_network - -logging.basicConfig( - format='[%(levelname)s %(asctime)s %(filename)s:%(lineno)d] %(message)s') - - -class DeepSpeech2Model(object): - """DeepSpeech2Model class. - - :param vocab_size: Decoding vocabulary size. - :type vocab_size: int - :param num_conv_layers: Number of stacking convolution layers. - :type num_conv_layers: int - :param num_rnn_layers: Number of stacking RNN layers. - :type num_rnn_layers: int - :param rnn_layer_size: RNN layer size (number of RNN cells). - :type rnn_layer_size: int - :param pretrained_model_path: Pretrained model path. If None, will train - from stratch. - :type pretrained_model_path: basestring|None - :param share_rnn_weights: Whether to share input-hidden weights between - forward and backward directional RNNs.Notice that - for GRU, weight sharing is not supported. - :type share_rnn_weights: bool - """ - - def __init__(self, vocab_size, num_conv_layers, num_rnn_layers, - rnn_layer_size, use_gru, pretrained_model_path, - share_rnn_weights): - self._create_network(vocab_size, num_conv_layers, num_rnn_layers, - rnn_layer_size, use_gru, share_rnn_weights) - self._create_parameters(pretrained_model_path) - self._inferer = None - self._loss_inferer = None - self._ext_scorer = None - self.logger = logging.getLogger("") - self.logger.setLevel(level=logging.INFO) - - def train(self, - train_batch_reader, - dev_batch_reader, - feeding_dict, - learning_rate, - gradient_clipping, - num_passes, - output_model_dir, - is_local=True, - num_iterations_print=100, - test_off=False): - """Train the model. - - :param train_batch_reader: Train data reader. - :type train_batch_reader: callable - :param dev_batch_reader: Validation data reader. - :type dev_batch_reader: callable - :param feeding_dict: Feeding is a map of field name and tuple index - of the data that reader returns. - :type feeding_dict: dict|list - :param learning_rate: Learning rate for ADAM optimizer. - :type learning_rate: float - :param gradient_clipping: Gradient clipping threshold. - :type gradient_clipping: float - :param num_passes: Number of training epochs. - :type num_passes: int - :param num_iterations_print: Number of training iterations for printing - a training loss. - :type rnn_iteratons_print: int - :param is_local: Set to False if running with pserver with multi-nodes. - :type is_local: bool - :param output_model_dir: Directory for saving the model (every pass). - :type output_model_dir: basestring - :param test_off: Turn off testing. - :type test_off: bool - """ - # prepare model output directory - if not os.path.exists(output_model_dir): - mkpath(output_model_dir) - - # prepare optimizer and trainer - optimizer = paddle.optimizer.Adam( - learning_rate=learning_rate, - gradient_clipping_threshold=gradient_clipping) - trainer = paddle.trainer.SGD( - cost=self._loss, - parameters=self._parameters, - update_equation=optimizer, - is_local=is_local) - - # create event handler - def event_handler(event): - global start_time, cost_sum, cost_counter - if isinstance(event, paddle.event.EndIteration): - cost_sum += event.cost - cost_counter += 1 - if (event.batch_id + 1) % num_iterations_print == 0: - output_model_path = os.path.join(output_model_dir, - "params.latest.tar.gz") - with gzip.open(output_model_path, 'w') as f: - trainer.save_parameter_to_tar(f) - print("\nPass: %d, Batch: %d, TrainCost: %f" % - (event.pass_id, event.batch_id + 1, - cost_sum / cost_counter)) - cost_sum, cost_counter = 0.0, 0 - else: - sys.stdout.write('.') - sys.stdout.flush() - if isinstance(event, paddle.event.BeginPass): - start_time = time.time() - cost_sum, cost_counter = 0.0, 0 - if isinstance(event, paddle.event.EndPass): - if test_off: - print("\n------- Time: %d sec, Pass: %d" % - (time.time() - start_time, event.pass_id)) - else: - result = trainer.test( - reader=dev_batch_reader, feeding=feeding_dict) - print( - "\n------- Time: %d sec, Pass: %d, " - "ValidationCost: %s" % - (time.time() - start_time, event.pass_id, result.cost)) - output_model_path = os.path.join( - output_model_dir, "params.pass-%d.tar.gz" % event.pass_id) - with gzip.open(output_model_path, 'w') as f: - trainer.save_parameter_to_tar(f) - - # run train - trainer.train( - reader=train_batch_reader, - event_handler=event_handler, - num_passes=num_passes, - feeding=feeding_dict) - - def infer_loss_batch(self, infer_data): - """Model inference. Infer the ctc loss for a batch of speech - utterances. - - :param infer_data: List of utterances to infer, with each utterance a - tuple of audio features and transcription text (empty - string). - :type infer_data: list - :return: List of ctc loss. - :rtype: List of float - """ - # define inferer - if self._loss_inferer == None: - self._loss_inferer = paddle.inference.Inference( - output_layer=self._loss, parameters=self._parameters) - # run inference - return self._loss_inferer.infer(input=infer_data) - - def infer_batch(self, infer_data, decoding_method, beam_alpha, beam_beta, - beam_size, cutoff_prob, cutoff_top_n, vocab_list, - language_model_path, num_processes, feeding_dict): - """Model inference. Infer the transcription for a batch of speech - utterances. - - :param infer_data: List of utterances to infer, with each utterance - consisting of a tuple of audio features and - transcription text (empty string). - :type infer_data: list - :param decoding_method: Decoding method name, 'ctc_greedy' or - 'ctc_beam_search'. - :param decoding_method: string - :param beam_alpha: Parameter associated with language model. - :type beam_alpha: float - :param beam_beta: Parameter associated with word count. - :type beam_beta: float - :param beam_size: Width for Beam search. - :type beam_size: int - :param cutoff_prob: Cutoff probability in pruning, - default 1.0, no pruning. - :type cutoff_prob: float - :param cutoff_top_n: Cutoff number in pruning, only top cutoff_top_n - characters with highest probs in vocabulary will be - used in beam search, default 40. - :type cutoff_top_n: int - :param vocab_list: List of tokens in the vocabulary, for decoding. - :type vocab_list: list - :param language_model_path: Filepath for language model. - :type language_model_path: basestring|None - :param num_processes: Number of processes (CPU) for decoder. - :type num_processes: int - :param feeding_dict: Feeding is a map of field name and tuple index - of the data that reader returns. - :type feeding_dict: dict|list - :return: List of transcription texts. - :rtype: List of basestring - """ - # define inferer - if self._inferer == None: - self._inferer = paddle.inference.Inference( - output_layer=self._log_probs, parameters=self._parameters) - # run inference - infer_results = self._inferer.infer( - input=infer_data, feeding=feeding_dict) - start_pos = [0] * (len(infer_data) + 1) - for i in xrange(len(infer_data)): - start_pos[i + 1] = start_pos[i] + infer_data[i][3][0] - probs_split = [ - infer_results[start_pos[i]:start_pos[i + 1]] - for i in xrange(0, len(infer_data)) - ] - # run decoder - results = [] - if decoding_method == "ctc_greedy": - # best path decode - for i, probs in enumerate(probs_split): - output_transcription = ctc_greedy_decoder( - probs_seq=probs, vocabulary=vocab_list) - results.append(output_transcription) - elif decoding_method == "ctc_beam_search": - # initialize external scorer - if self._ext_scorer == None: - self._loaded_lm_path = language_model_path - self.logger.info("begin to initialize the external scorer " - "for decoding") - self._ext_scorer = Scorer(beam_alpha, beam_beta, - language_model_path, vocab_list) - - lm_char_based = self._ext_scorer.is_character_based() - lm_max_order = self._ext_scorer.get_max_order() - lm_dict_size = self._ext_scorer.get_dict_size() - self.logger.info("language model: " - "is_character_based = %d," % lm_char_based + - " max_order = %d," % lm_max_order + - " dict_size = %d" % lm_dict_size) - self.logger.info("end initializing scorer. Start decoding ...") - else: - self._ext_scorer.reset_params(beam_alpha, beam_beta) - assert self._loaded_lm_path == language_model_path - # beam search decode - num_processes = min(num_processes, len(probs_split)) - beam_search_results = ctc_beam_search_decoder_batch( - probs_split=probs_split, - vocabulary=vocab_list, - beam_size=beam_size, - num_processes=num_processes, - ext_scoring_func=self._ext_scorer, - cutoff_prob=cutoff_prob, - cutoff_top_n=cutoff_top_n) - - results = [result[0][1] for result in beam_search_results] - else: - raise ValueError("Decoding method [%s] is not supported." % - decoding_method) - return results - - def _create_parameters(self, model_path=None): - """Load or create model parameters.""" - if model_path is None: - self._parameters = paddle.parameters.create(self._loss) - else: - self._parameters = paddle.parameters.Parameters.from_tar( - gzip.open(model_path)) - - def _create_network(self, vocab_size, num_conv_layers, num_rnn_layers, - rnn_layer_size, use_gru, share_rnn_weights): - """Create data layers and model network.""" - # paddle.data_type.dense_array is used for variable batch input. - # The size 161 * 161 is only an placeholder value and the real shape - # of input batch data will be induced during training. - audio_data = paddle.layer.data( - name="audio_spectrogram", - type=paddle.data_type.dense_array(161 * 161)) - text_data = paddle.layer.data( - name="transcript_text", - type=paddle.data_type.integer_value_sequence(vocab_size)) - seq_offset_data = paddle.layer.data( - name='sequence_offset', - type=paddle.data_type.integer_value_sequence(1)) - seq_len_data = paddle.layer.data( - name='sequence_length', - type=paddle.data_type.integer_value_sequence(1)) - index_range_datas = [] - for i in xrange(num_rnn_layers): - index_range_datas.append( - paddle.layer.data( - name='conv%d_index_range' % i, - type=paddle.data_type.dense_vector(6))) - - self._log_probs, self._loss = deep_speech_v2_network( - audio_data=audio_data, - text_data=text_data, - seq_offset_data=seq_offset_data, - seq_len_data=seq_len_data, - index_range_datas=index_range_datas, - dict_size=vocab_size, - num_conv_layers=num_conv_layers, - num_rnn_layers=num_rnn_layers, - rnn_size=rnn_layer_size, - use_gru=use_gru, - share_rnn_weights=share_rnn_weights) diff --git a/deep_speech_2/model_utils/network.py b/deep_speech_2/model_utils/network.py deleted file mode 100644 index 7b4b8ab20c702e96daaa155877d3e0c9eeb4a17b..0000000000000000000000000000000000000000 --- a/deep_speech_2/model_utils/network.py +++ /dev/null @@ -1,302 +0,0 @@ -"""Contains DeepSpeech2 layers and networks.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import paddle.v2 as paddle - - -def conv_bn_layer(input, filter_size, num_channels_in, num_channels_out, stride, - padding, act, index_range_data): - """Convolution layer with batch normalization. - - :param input: Input layer. - :type input: LayerOutput - :param filter_size: The x dimension of a filter kernel. Or input a tuple for - two image dimension. - :type filter_size: int|tuple|list - :param num_channels_in: Number of input channels. - :type num_channels_in: int - :type num_channels_out: Number of output channels. - :type num_channels_in: out - :param padding: The x dimension of the padding. Or input a tuple for two - image dimension. - :type padding: int|tuple|list - :param act: Activation type. - :type act: BaseActivation - :param index_range_data: Index range to indicate sub region. - :type index_range_data: LayerOutput - :return: Batch norm layer after convolution layer. - :rtype: LayerOutput - """ - conv_layer = paddle.layer.img_conv( - input=input, - filter_size=filter_size, - num_channels=num_channels_in, - num_filters=num_channels_out, - stride=stride, - padding=padding, - act=paddle.activation.Linear(), - bias_attr=False) - batch_norm = paddle.layer.batch_norm(input=conv_layer, act=act) - # reset padding part to 0 - scale_sub_region = paddle.layer.scale_sub_region( - batch_norm, index_range_data, value=0.0) - return scale_sub_region - - -def bidirectional_simple_rnn_bn_layer(name, input, size, act, share_weights): - """Bidirectonal simple rnn layer with sequence-wise batch normalization. - The batch normalization is only performed on input-state weights. - - :param name: Name of the layer. - :type name: string - :param input: Input layer. - :type input: LayerOutput - :param size: Number of RNN cells. - :type size: int - :param act: Activation type. - :type act: BaseActivation - :param share_weights: Whether to share input-hidden weights between - forward and backward directional RNNs. - :type share_weights: bool - :return: Bidirectional simple rnn layer. - :rtype: LayerOutput - """ - if share_weights: - # input-hidden weights shared between bi-direcitonal rnn. - input_proj = paddle.layer.fc( - input=input, - size=size, - act=paddle.activation.Linear(), - bias_attr=False) - # batch norm is only performed on input-state projection - input_proj_bn = paddle.layer.batch_norm( - input=input_proj, act=paddle.activation.Linear()) - # forward and backward in time - forward_simple_rnn = paddle.layer.recurrent( - input=input_proj_bn, act=act, reverse=False) - backward_simple_rnn = paddle.layer.recurrent( - input=input_proj_bn, act=act, reverse=True) - - else: - input_proj_forward = paddle.layer.fc( - input=input, - size=size, - act=paddle.activation.Linear(), - bias_attr=False) - input_proj_backward = paddle.layer.fc( - input=input, - size=size, - act=paddle.activation.Linear(), - bias_attr=False) - # batch norm is only performed on input-state projection - input_proj_bn_forward = paddle.layer.batch_norm( - input=input_proj_forward, act=paddle.activation.Linear()) - input_proj_bn_backward = paddle.layer.batch_norm( - input=input_proj_backward, act=paddle.activation.Linear()) - # forward and backward in time - forward_simple_rnn = paddle.layer.recurrent( - input=input_proj_bn_forward, act=act, reverse=False) - backward_simple_rnn = paddle.layer.recurrent( - input=input_proj_bn_backward, act=act, reverse=True) - - return paddle.layer.concat(input=[forward_simple_rnn, backward_simple_rnn]) - - -def bidirectional_gru_bn_layer(name, input, size, act): - """Bidirectonal gru layer with sequence-wise batch normalization. - The batch normalization is only performed on input-state weights. - - :param name: Name of the layer. - :type name: string - :param input: Input layer. - :type input: LayerOutput - :param size: Number of RNN cells. - :type size: int - :param act: Activation type. - :type act: BaseActivation - :return: Bidirectional simple rnn layer. - :rtype: LayerOutput - """ - input_proj_forward = paddle.layer.fc( - input=input, - size=size * 3, - act=paddle.activation.Linear(), - bias_attr=False) - input_proj_backward = paddle.layer.fc( - input=input, - size=size * 3, - act=paddle.activation.Linear(), - bias_attr=False) - # batch norm is only performed on input-related projections - input_proj_bn_forward = paddle.layer.batch_norm( - input=input_proj_forward, act=paddle.activation.Linear()) - input_proj_bn_backward = paddle.layer.batch_norm( - input=input_proj_backward, act=paddle.activation.Linear()) - # forward and backward in time - forward_gru = paddle.layer.grumemory( - input=input_proj_bn_forward, act=act, reverse=False) - backward_gru = paddle.layer.grumemory( - input=input_proj_bn_backward, act=act, reverse=True) - return paddle.layer.concat(input=[forward_gru, backward_gru]) - - -def conv_group(input, num_stacks, index_range_datas): - """Convolution group with stacked convolution layers. - - :param input: Input layer. - :type input: LayerOutput - :param num_stacks: Number of stacked convolution layers. - :type num_stacks: int - :param index_range_datas: Index ranges for each convolution layer. - :type index_range_datas: tuple|list - :return: Output layer of the convolution group. - :rtype: LayerOutput - """ - conv = conv_bn_layer( - input=input, - filter_size=(11, 41), - num_channels_in=1, - num_channels_out=32, - stride=(3, 2), - padding=(5, 20), - act=paddle.activation.BRelu(), - index_range_data=index_range_datas[0]) - for i in xrange(num_stacks - 1): - conv = conv_bn_layer( - input=conv, - filter_size=(11, 21), - num_channels_in=32, - num_channels_out=32, - stride=(1, 2), - padding=(5, 10), - act=paddle.activation.BRelu(), - index_range_data=index_range_datas[i + 1]) - output_num_channels = 32 - output_height = 160 // pow(2, num_stacks) + 1 - return conv, output_num_channels, output_height - - -def rnn_group(input, size, num_stacks, use_gru, share_rnn_weights): - """RNN group with stacked bidirectional simple RNN layers. - - :param input: Input layer. - :type input: LayerOutput - :param size: Number of RNN cells in each layer. - :type size: int - :param num_stacks: Number of stacked rnn layers. - :type num_stacks: int - :param use_gru: Use gru if set True. Use simple rnn if set False. - :type use_gru: bool - :param share_rnn_weights: Whether to share input-hidden weights between - forward and backward directional RNNs. - It is only available when use_gru=False. - :type share_weights: bool - :return: Output layer of the RNN group. - :rtype: LayerOutput - """ - output = input - for i in xrange(num_stacks): - if use_gru: - output = bidirectional_gru_bn_layer( - name=str(i), - input=output, - size=size, - act=paddle.activation.Relu()) - # BRelu does not support hppl, need to add later. Use Relu instead. - else: - output = bidirectional_simple_rnn_bn_layer( - name=str(i), - input=output, - size=size, - act=paddle.activation.BRelu(), - share_weights=share_rnn_weights) - return output - - -def deep_speech_v2_network(audio_data, - text_data, - seq_offset_data, - seq_len_data, - index_range_datas, - dict_size, - num_conv_layers=2, - num_rnn_layers=3, - rnn_size=256, - use_gru=False, - share_rnn_weights=True): - """The DeepSpeech2 network structure. - - :param audio_data: Audio spectrogram data layer. - :type audio_data: LayerOutput - :param text_data: Transcription text data layer. - :type text_data: LayerOutput - :param seq_offset_data: Sequence offset data layer. - :type seq_offset_data: LayerOutput - :param seq_len_data: Valid sequence length data layer. - :type seq_len_data: LayerOutput - :param index_range_datas: Index ranges data layers. - :type index_range_datas: tuple|list - :param dict_size: Dictionary size for tokenized transcription. - :type dict_size: int - :param num_conv_layers: Number of stacking convolution layers. - :type num_conv_layers: int - :param num_rnn_layers: Number of stacking RNN layers. - :type num_rnn_layers: int - :param rnn_size: RNN layer size (number of RNN cells). - :type rnn_size: int - :param use_gru: Use gru if set True. Use simple rnn if set False. - :type use_gru: bool - :param share_rnn_weights: Whether to share input-hidden weights between - forward and backward direction RNNs. - It is only available when use_gru=False. - :type share_weights: bool - :return: A tuple of an output unnormalized log probability layer ( - before softmax) and a ctc cost layer. - :rtype: tuple of LayerOutput - """ - # convolution group - conv_group_output, conv_group_num_channels, conv_group_height = conv_group( - input=audio_data, - num_stacks=num_conv_layers, - index_range_datas=index_range_datas) - # convert data form convolution feature map to sequence of vectors - conv2seq = paddle.layer.block_expand( - input=conv_group_output, - num_channels=conv_group_num_channels, - stride_x=1, - stride_y=1, - block_x=1, - block_y=conv_group_height) - # remove padding part - remove_padding_data = paddle.layer.sub_seq( - input=conv2seq, - offsets=seq_offset_data, - sizes=seq_len_data, - act=paddle.activation.Linear(), - bias_attr=False) - # rnn group - rnn_group_output = rnn_group( - input=remove_padding_data, - size=rnn_size, - num_stacks=num_rnn_layers, - use_gru=use_gru, - share_rnn_weights=share_rnn_weights) - fc = paddle.layer.fc( - input=rnn_group_output, - size=dict_size + 1, - act=paddle.activation.Linear(), - bias_attr=True) - # probability distribution with softmax - log_probs = paddle.layer.mixed( - input=paddle.layer.identity_projection(input=fc), - act=paddle.activation.Softmax()) - # ctc cost - ctc_loss = paddle.layer.warp_ctc( - input=fc, - label=text_data, - size=dict_size + 1, - blank=dict_size, - norm_by_times=True) - return log_probs, ctc_loss diff --git a/deep_speech_2/models/aishell/download_model.sh b/deep_speech_2/models/aishell/download_model.sh deleted file mode 100644 index 19aec554e209690cde4151b7f168bcf542c22224..0000000000000000000000000000000000000000 --- a/deep_speech_2/models/aishell/download_model.sh +++ /dev/null @@ -1,19 +0,0 @@ -#! /usr/bin/env bash - -. ../../utils/utility.sh - -URL='http://cloud.dlnel.org/filepub/?uuid=6c83b9d8-3255-4adf-9726-0fe0be3d0274' -MD5=28521a58552885a81cf92a1e9b133a71 -TARGET=./aishell_model.tar.gz - - -echo "Download Aishell model ..." -download $URL $MD5 $TARGET -if [ $? -ne 0 ]; then - echo "Fail to download Aishell model!" - exit 1 -fi -tar -zxvf $TARGET - - -exit 0 diff --git a/deep_speech_2/models/baidu_en8k/download_model.sh b/deep_speech_2/models/baidu_en8k/download_model.sh deleted file mode 100644 index 7d92fd52eff30d5877b2393a61b6128b7b1a1321..0000000000000000000000000000000000000000 --- a/deep_speech_2/models/baidu_en8k/download_model.sh +++ /dev/null @@ -1,19 +0,0 @@ -#! /usr/bin/env bash - -. ../../utils/utility.sh - -URL='To-be-added' -MD5=a19d40cb3b558eb696c44d883f32cfda -TARGET=./baidu_en8k_model.tar.gz - - -echo "Download BaiduEn8k model ..." -download $URL $MD5 $TARGET -if [ $? -ne 0 ]; then - echo "Fail to download BaiduEn8k model!" - exit 1 -fi -tar -zxvf $TARGET - - -exit 0 diff --git a/deep_speech_2/models/librispeech/download_model.sh b/deep_speech_2/models/librispeech/download_model.sh deleted file mode 100644 index 9c0ec2783bb08392c93f5959274b84355fc3a58c..0000000000000000000000000000000000000000 --- a/deep_speech_2/models/librispeech/download_model.sh +++ /dev/null @@ -1,19 +0,0 @@ -#! /usr/bin/env bash - -. ../../utils/utility.sh - -URL='http://cloud.dlnel.org/filepub/?uuid=6020a634-5399-4423-b021-c5ed32680fff' -MD5=2ef08f8b608a7c555592161fc14d81a6 -TARGET=./librispeech_model.tar.gz - - -echo "Download LibriSpeech model ..." -download $URL $MD5 $TARGET -if [ $? -ne 0 ]; then - echo "Fail to download LibriSpeech model!" - exit 1 -fi -tar -zxvf $TARGET - - -exit 0 diff --git a/deep_speech_2/models/lm/download_lm_ch.sh b/deep_speech_2/models/lm/download_lm_ch.sh deleted file mode 100644 index c719f9a9f26368c17c8ef8651f36dfec194da0a2..0000000000000000000000000000000000000000 --- a/deep_speech_2/models/lm/download_lm_ch.sh +++ /dev/null @@ -1,18 +0,0 @@ -#! /usr/bin/env bash - -. ../../utils/utility.sh - -URL=http://cloud.dlnel.org/filepub/?uuid=d21861e4-4ed6-45bb-ad8e-ae417a43195e -MD5="29e02312deb2e59b3c8686c7966d4fe3" -TARGET=./zh_giga.no_cna_cmn.prune01244.klm - - -echo "Download language model ..." -download $URL $MD5 $TARGET -if [ $? -ne 0 ]; then - echo "Fail to download the language model!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/models/lm/download_lm_en.sh b/deep_speech_2/models/lm/download_lm_en.sh deleted file mode 100644 index d131636e80b0907a94712eef6baf72afadb05477..0000000000000000000000000000000000000000 --- a/deep_speech_2/models/lm/download_lm_en.sh +++ /dev/null @@ -1,18 +0,0 @@ -#! /usr/bin/env bash - -. ../../utils/utility.sh - -URL=http://paddlepaddle.bj.bcebos.com/model_zoo/speech/common_crawl_00.prune01111.trie.klm -MD5="099a601759d467cd0a8523ff939819c5" -TARGET=./common_crawl_00.prune01111.trie.klm - - -echo "Download language model ..." -download $URL $MD5 $TARGET -if [ $? -ne 0 ]; then - echo "Fail to download the language model!" - exit 1 -fi - - -exit 0 diff --git a/deep_speech_2/requirements.txt b/deep_speech_2/requirements.txt deleted file mode 100644 index e104f633c792ee9be4f2864e99aea64924a1e2fb..0000000000000000000000000000000000000000 --- a/deep_speech_2/requirements.txt +++ /dev/null @@ -1,4 +0,0 @@ -scipy==0.13.1 -resampy==0.1.5 -SoundFile==0.9.0.post1 -python_speech_features diff --git a/deep_speech_2/setup.sh b/deep_speech_2/setup.sh deleted file mode 100644 index 7c40415db324deed8d2279f5d31287bb21449f55..0000000000000000000000000000000000000000 --- a/deep_speech_2/setup.sh +++ /dev/null @@ -1,38 +0,0 @@ -#! /usr/bin/env bash - -# install python dependencies -if [ -f "requirements.txt" ]; then - pip install -r requirements.txt -fi -if [ $? != 0 ]; then - echo "Install python dependencies failed !!!" - exit 1 -fi - -# install package libsndfile -python -c "import soundfile" -if [ $? != 0 ]; then - echo "Install package libsndfile into default system path." - wget "http://www.mega-nerd.com/libsndfile/files/libsndfile-1.0.28.tar.gz" - if [ $? != 0 ]; then - echo "Download libsndfile-1.0.28.tar.gz failed !!!" - exit 1 - fi - tar -zxvf libsndfile-1.0.28.tar.gz - cd libsndfile-1.0.28 - ./configure > /dev/null && make > /dev/null && make install > /dev/null - cd .. - rm -rf libsndfile-1.0.28 - rm libsndfile-1.0.28.tar.gz -fi - -# install decoders -python -c "import swig_decoders" -if [ $? != 0 ]; then - cd decoders/swig > /dev/null - sh setup.sh - cd - > /dev/null -fi - - -echo "Install all dependencies successfully." diff --git a/deep_speech_2/test.py b/deep_speech_2/test.py deleted file mode 100644 index 53f7e17af3d9ceb854301e67fb0d4698f6d41d39..0000000000000000000000000000000000000000 --- a/deep_speech_2/test.py +++ /dev/null @@ -1,129 +0,0 @@ -"""Evaluation for DeepSpeech2 model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import argparse -import functools -import paddle.v2 as paddle -from data_utils.data import DataGenerator -from model_utils.model import DeepSpeech2Model -from utils.error_rate import wer, cer -from utils.utility import add_arguments, print_arguments - -parser = argparse.ArgumentParser(description=__doc__) -add_arg = functools.partial(add_arguments, argparser=parser) -# yapf: disable -add_arg('batch_size', int, 128, "Minibatch size.") -add_arg('trainer_count', int, 8, "# of Trainers (CPUs or GPUs).") -add_arg('beam_size', int, 500, "Beam search width.") -add_arg('num_proc_bsearch', int, 8, "# of CPUs for beam search.") -add_arg('num_proc_data', int, 8, "# of CPUs for data preprocessing.") -add_arg('num_conv_layers', int, 2, "# of convolution layers.") -add_arg('num_rnn_layers', int, 3, "# of recurrent layers.") -add_arg('rnn_layer_size', int, 2048, "# of recurrent cells per layer.") -add_arg('alpha', float, 2.15, "Coef of LM for beam search.") -add_arg('beta', float, 0.35, "Coef of WC for beam search.") -add_arg('cutoff_prob', float, 1.0, "Cutoff probability for pruning.") -add_arg('cutoff_top_n', int, 40, "Cutoff number for pruning.") -add_arg('use_gru', bool, False, "Use GRUs instead of simple RNNs.") -add_arg('use_gpu', bool, True, "Use GPU or not.") -add_arg('share_rnn_weights',bool, True, "Share input-hidden weights across " - "bi-directional RNNs. Not for GRU.") -add_arg('test_manifest', str, - 'data/librispeech/manifest.test-clean', - "Filepath of manifest to evaluate.") -add_arg('mean_std_path', str, - 'data/librispeech/mean_std.npz', - "Filepath of normalizer's mean & std.") -add_arg('vocab_path', str, - 'data/librispeech/vocab.txt', - "Filepath of vocabulary.") -add_arg('model_path', str, - './checkpoints/libri/params.latest.tar.gz', - "If None, the training starts from scratch, " - "otherwise, it resumes from the pre-trained model.") -add_arg('lang_model_path', str, - 'models/lm/common_crawl_00.prune01111.trie.klm', - "Filepath for language model.") -add_arg('decoding_method', str, - 'ctc_beam_search', - "Decoding method. Options: ctc_beam_search, ctc_greedy", - choices = ['ctc_beam_search', 'ctc_greedy']) -add_arg('error_rate_type', str, - 'wer', - "Error rate type for evaluation.", - choices=['wer', 'cer']) -add_arg('specgram_type', str, - 'linear', - "Audio feature type. Options: linear, mfcc.", - choices=['linear', 'mfcc']) -# yapf: disable -args = parser.parse_args() - - -def evaluate(): - """Evaluate on whole test data for DeepSpeech2.""" - data_generator = DataGenerator( - vocab_filepath=args.vocab_path, - mean_std_filepath=args.mean_std_path, - augmentation_config='{}', - specgram_type=args.specgram_type, - num_threads=args.num_proc_data, - keep_transcription_text=True, - num_conv_layers=args.num_conv_layers) - batch_reader = data_generator.batch_reader_creator( - manifest_path=args.test_manifest, - batch_size=args.batch_size, - min_batch_size=1, - sortagrad=False, - shuffle_method=None) - - ds2_model = DeepSpeech2Model( - vocab_size=data_generator.vocab_size, - num_conv_layers=args.num_conv_layers, - num_rnn_layers=args.num_rnn_layers, - rnn_layer_size=args.rnn_layer_size, - use_gru=args.use_gru, - pretrained_model_path=args.model_path, - share_rnn_weights=args.share_rnn_weights) - - # decoders only accept string encoded in utf-8 - vocab_list = [chars.encode("utf-8") for chars in data_generator.vocab_list] - - error_rate_func = cer if args.error_rate_type == 'cer' else wer - error_sum, num_ins = 0.0, 0 - for infer_data in batch_reader(): - result_transcripts = ds2_model.infer_batch( - infer_data=infer_data, - decoding_method=args.decoding_method, - beam_alpha=args.alpha, - beam_beta=args.beta, - beam_size=args.beam_size, - cutoff_prob=args.cutoff_prob, - cutoff_top_n=args.cutoff_top_n, - vocab_list=vocab_list, - language_model_path=args.lang_model_path, - num_processes=args.num_proc_bsearch, - feeding_dict=data_generator.feeding) - target_transcripts = [data[1] for data in infer_data] - for target, result in zip(target_transcripts, result_transcripts): - error_sum += error_rate_func(target, result) - num_ins += 1 - print("Error rate [%s] (%d/?) = %f" % - (args.error_rate_type, num_ins, error_sum / num_ins)) - print("Final error rate [%s] (%d/%d) = %f" % - (args.error_rate_type, num_ins, num_ins, error_sum / num_ins)) - - ds2_model.logger.info("finish evaluation") - -def main(): - print_arguments(args) - paddle.init(use_gpu=args.use_gpu, - rnn_use_batch=True, - trainer_count=args.trainer_count) - evaluate() - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/tools/_init_paths.py b/deep_speech_2/tools/_init_paths.py deleted file mode 100644 index ddabb535be682d95c3c8b73003ea30eed06ca0b0..0000000000000000000000000000000000000000 --- a/deep_speech_2/tools/_init_paths.py +++ /dev/null @@ -1,19 +0,0 @@ -"""Set up paths for DS2""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import os.path -import sys - - -def add_path(path): - if path not in sys.path: - sys.path.insert(0, path) - - -this_dir = os.path.dirname(__file__) - -# Add project path to PYTHONPATH -proj_path = os.path.join(this_dir, '..') -add_path(proj_path) diff --git a/deep_speech_2/tools/build_vocab.py b/deep_speech_2/tools/build_vocab.py deleted file mode 100644 index e167e92adf36a75972f9695e76b29cc2adfb4f77..0000000000000000000000000000000000000000 --- a/deep_speech_2/tools/build_vocab.py +++ /dev/null @@ -1,58 +0,0 @@ -"""Build vocabulary from manifest files. - -Each item in vocabulary file is a character. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import argparse -import functools -import codecs -import json -from collections import Counter -import os.path -import _init_paths -from data_utils.utility import read_manifest -from utils.utility import add_arguments, print_arguments - -parser = argparse.ArgumentParser(description=__doc__) -add_arg = functools.partial(add_arguments, argparser=parser) -# yapf: disable -add_arg('count_threshold', int, 0, "Truncation threshold for char counts.") -add_arg('vocab_path', str, - 'data/librispeech/vocab.txt', - "Filepath to write the vocabulary.") -add_arg('manifest_paths', str, - None, - "Filepaths of manifests for building vocabulary. " - "You can provide multiple manifest files.", - nargs='+', - required=True) -# yapf: disable -args = parser.parse_args() - - -def count_manifest(counter, manifest_path): - manifest_jsons = read_manifest(manifest_path) - for line_json in manifest_jsons: - for char in line_json['text']: - counter.update(char) - - -def main(): - print_arguments(args) - - counter = Counter() - for manifest_path in args.manifest_paths: - count_manifest(counter, manifest_path) - - count_sorted = sorted(counter.items(), key=lambda x: x[1], reverse=True) - with codecs.open(args.vocab_path, 'w', 'utf-8') as fout: - for char, count in count_sorted: - if count < args.count_threshold: break - fout.write(char + '\n') - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/tools/compute_mean_std.py b/deep_speech_2/tools/compute_mean_std.py deleted file mode 100644 index 11aa856d709c4bf4ace1ec7f842ff16b3bc24ad0..0000000000000000000000000000000000000000 --- a/deep_speech_2/tools/compute_mean_std.py +++ /dev/null @@ -1,51 +0,0 @@ -"""Compute mean and std for feature normalizer, and save to file.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import argparse -import functools -import _init_paths -from data_utils.normalizer import FeatureNormalizer -from data_utils.augmentor.augmentation import AugmentationPipeline -from data_utils.featurizer.audio_featurizer import AudioFeaturizer -from utils.utility import add_arguments, print_arguments - -parser = argparse.ArgumentParser(description=__doc__) -add_arg = functools.partial(add_arguments, argparser=parser) -# yapf: disable -add_arg('num_samples', int, 2000, "# of samples to for statistics.") -add_arg('specgram_type', str, - 'linear', - "Audio feature type. Options: linear, mfcc.", - choices=['linear', 'mfcc']) -add_arg('manifest_path', str, - 'data/librispeech/manifest.train', - "Filepath of manifest to compute normalizer's mean and stddev.") -add_arg('output_path', str, - 'data/librispeech/mean_std.npz', - "Filepath of write mean and stddev to (.npz).") -# yapf: disable -args = parser.parse_args() - - -def main(): - print_arguments(args) - - augmentation_pipeline = AugmentationPipeline('{}') - audio_featurizer = AudioFeaturizer(specgram_type=args.specgram_type) - - def augment_and_featurize(audio_segment): - augmentation_pipeline.transform_audio(audio_segment) - return audio_featurizer.featurize(audio_segment) - - normalizer = FeatureNormalizer( - mean_std_filepath=None, - manifest_path=args.manifest_path, - featurize_func=augment_and_featurize, - num_samples=args.num_samples) - normalizer.write_to_file(args.output_path) - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/tools/profile.sh b/deep_speech_2/tools/profile.sh deleted file mode 100644 index 19abe7ede17b213806214245c6aad912577e37ce..0000000000000000000000000000000000000000 --- a/deep_speech_2/tools/profile.sh +++ /dev/null @@ -1,30 +0,0 @@ -#! /usr/bin/env bash - -BATCH_SIZE_PER_GPU=64 -MIN_DURATION=6.0 -MAX_DURATION=7.0 - -function join_by { local IFS="$1"; shift; echo "$*"; } - -for NUM_GPUS in 16 8 4 2 1 -do - DEVICES=$(join_by , $(seq 0 $(($NUM_GPUS-1)))) - BATCH_SIZE=$(($BATCH_SIZE_PER_GPU * $NUM_GPUS)) - - CUDA_VISIBLE_DEVICES=$DEVICES \ - python train.py \ - --batch_size=$BATCH_SIZE \ - --num_passes=1 \ - --test_off=True \ - --trainer_count=$NUM_GPUS \ - --min_duration=$MIN_DURATION \ - --max_duration=$MAX_DURATION > tmp.log 2>&1 - - if [ $? -ne 0 ];then - exit 1 - fi - - cat tmp.log | grep "Time" | awk '{print "GPU Num: " "'"$NUM_GPUS"'" " Time: "$3}' - - rm tmp.log -done diff --git a/deep_speech_2/tools/tune.py b/deep_speech_2/tools/tune.py deleted file mode 100644 index 47abf141349dba7a99b154e4d919b75b43a6ad59..0000000000000000000000000000000000000000 --- a/deep_speech_2/tools/tune.py +++ /dev/null @@ -1,244 +0,0 @@ -"""Beam search parameters tuning for DeepSpeech2 model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import sys -import os -import numpy as np -import argparse -import functools -import gzip -import logging -import paddle.v2 as paddle -import _init_paths -from data_utils.data import DataGenerator -from decoders.swig_wrapper import Scorer -from decoders.swig_wrapper import ctc_beam_search_decoder_batch -from model_utils.model import deep_speech_v2_network -from utils.error_rate import wer, cer -from utils.utility import add_arguments, print_arguments - -parser = argparse.ArgumentParser(description=__doc__) -add_arg = functools.partial(add_arguments, argparser=parser) -# yapf: disable -add_arg('num_batches', int, -1, "# of batches tuning on. " - "Default -1, on whole dev set.") -add_arg('batch_size', int, 256, "# of samples per batch.") -add_arg('trainer_count', int, 8, "# of Trainers (CPUs or GPUs).") -add_arg('beam_size', int, 500, "Beam search width.") -add_arg('num_proc_bsearch', int, 8, "# of CPUs for beam search.") -add_arg('num_proc_data', int, 8, "# of CPUs for data preprocessing.") -add_arg('num_conv_layers', int, 2, "# of convolution layers.") -add_arg('num_rnn_layers', int, 3, "# of recurrent layers.") -add_arg('rnn_layer_size', int, 2048, "# of recurrent cells per layer.") -add_arg('num_alphas', int, 45, "# of alpha candidates for tuning.") -add_arg('num_betas', int, 8, "# of beta candidates for tuning.") -add_arg('alpha_from', float, 1.0, "Where alpha starts tuning from.") -add_arg('alpha_to', float, 3.2, "Where alpha ends tuning with.") -add_arg('beta_from', float, 0.1, "Where beta starts tuning from.") -add_arg('beta_to', float, 0.45, "Where beta ends tuning with.") -add_arg('cutoff_prob', float, 1.0, "Cutoff probability for pruning.") -add_arg('cutoff_top_n', int, 40, "Cutoff number for pruning.") -add_arg('use_gru', bool, False, "Use GRUs instead of simple RNNs.") -add_arg('use_gpu', bool, True, "Use GPU or not.") -add_arg('share_rnn_weights',bool, True, "Share input-hidden weights across " - "bi-directional RNNs. Not for GRU.") -add_arg('tune_manifest', str, - 'data/librispeech/manifest.dev-clean', - "Filepath of manifest to tune.") -add_arg('mean_std_path', str, - 'data/librispeech/mean_std.npz', - "Filepath of normalizer's mean & std.") -add_arg('vocab_path', str, - 'data/librispeech/vocab.txt', - "Filepath of vocabulary.") -add_arg('lang_model_path', str, - 'models/lm/common_crawl_00.prune01111.trie.klm', - "Filepath for language model.") -add_arg('model_path', str, - './checkpoints/libri/params.latest.tar.gz', - "If None, the training starts from scratch, " - "otherwise, it resumes from the pre-trained model.") -add_arg('error_rate_type', str, - 'wer', - "Error rate type for evaluation.", - choices=['wer', 'cer']) -add_arg('specgram_type', str, - 'linear', - "Audio feature type. Options: linear, mfcc.", - choices=['linear', 'mfcc']) -# yapf: disable -args = parser.parse_args() - - -logging.basicConfig( - format='[%(levelname)s %(asctime)s %(filename)s:%(lineno)d] %(message)s') - -def tune(): - """Tune parameters alpha and beta incrementally.""" - if not args.num_alphas >= 0: - raise ValueError("num_alphas must be non-negative!") - if not args.num_betas >= 0: - raise ValueError("num_betas must be non-negative!") - - data_generator = DataGenerator( - vocab_filepath=args.vocab_path, - mean_std_filepath=args.mean_std_path, - augmentation_config='{}', - specgram_type=args.specgram_type, - num_threads=args.num_proc_data, - keep_transcription_text=True, - num_conv_layers=args.num_conv_layers) - - audio_data = paddle.layer.data( - name="audio_spectrogram", - type=paddle.data_type.dense_array(161 * 161)) - text_data = paddle.layer.data( - name="transcript_text", - type=paddle.data_type.integer_value_sequence(data_generator.vocab_size)) - seq_offset_data = paddle.layer.data( - name='sequence_offset', - type=paddle.data_type.integer_value_sequence(1)) - seq_len_data = paddle.layer.data( - name='sequence_length', - type=paddle.data_type.integer_value_sequence(1)) - index_range_datas = [] - for i in xrange(args.num_rnn_layers): - index_range_datas.append( - paddle.layer.data( - name='conv%d_index_range' % i, - type=paddle.data_type.dense_vector(6))) - - output_probs, _ = deep_speech_v2_network( - audio_data=audio_data, - text_data=text_data, - seq_offset_data=seq_offset_data, - seq_len_data=seq_len_data, - index_range_datas=index_range_datas, - dict_size=data_generator.vocab_size, - num_conv_layers=args.num_conv_layers, - num_rnn_layers=args.num_rnn_layers, - rnn_size=args.rnn_layer_size, - use_gru=args.use_gru, - share_rnn_weights=args.share_rnn_weights) - - batch_reader = data_generator.batch_reader_creator( - manifest_path=args.tune_manifest, - batch_size=args.batch_size, - sortagrad=False, - shuffle_method=None) - - # load parameters - if not os.path.isfile(args.model_path): - raise IOError("Invaid model path: %s" % args.model_path) - parameters = paddle.parameters.Parameters.from_tar( - gzip.open(args.model_path)) - - inferer = paddle.inference.Inference( - output_layer=output_probs, parameters=parameters) - # decoders only accept string encoded in utf-8 - vocab_list = [chars.encode("utf-8") for chars in data_generator.vocab_list] - - # init logger - logger = logging.getLogger("") - logger.setLevel(level=logging.INFO) - # init external scorer - logger.info("begin to initialize the external scorer for tuning") - if not os.path.isfile(args.lang_model_path): - raise IOError("Invaid language model path: %s" % args.lang_model_path) - ext_scorer = Scorer( - alpha=args.alpha_from, - beta=args.beta_from, - model_path=args.lang_model_path, - vocabulary=vocab_list) - logger.info("language model: " - "is_character_based = %d," % ext_scorer.is_character_based() + - " max_order = %d," % ext_scorer.get_max_order() + - " dict_size = %d" % ext_scorer.get_dict_size()) - logger.info("end initializing scorer. Start tuning ...") - - error_rate_func = cer if args.error_rate_type == 'cer' else wer - # create grid for search - cand_alphas = np.linspace(args.alpha_from, args.alpha_to, args.num_alphas) - cand_betas = np.linspace(args.beta_from, args.beta_to, args.num_betas) - params_grid = [(alpha, beta) for alpha in cand_alphas - for beta in cand_betas] - - err_sum = [0.0 for i in xrange(len(params_grid))] - err_ave = [0.0 for i in xrange(len(params_grid))] - num_ins, cur_batch = 0, 0 - ## incremental tuning parameters over multiple batches - for infer_data in batch_reader(): - if (args.num_batches >= 0) and (cur_batch >= args.num_batches): - break - infer_results = inferer.infer(input=infer_data, - feeding=data_generator.feeding) - start_pos = [0] * (len(infer_data) + 1) - for i in xrange(len(infer_data)): - start_pos[i + 1] = start_pos[i] + infer_data[i][3][0] - probs_split = [ - infer_results[start_pos[i]:start_pos[i + 1]] - for i in xrange(0, len(infer_data)) - ] - - target_transcripts = [ data[1] for data in infer_data ] - - num_ins += len(target_transcripts) - # grid search - for index, (alpha, beta) in enumerate(params_grid): - # reset alpha & beta - ext_scorer.reset_params(alpha, beta) - beam_search_results = ctc_beam_search_decoder_batch( - probs_split=probs_split, - vocabulary=vocab_list, - beam_size=args.beam_size, - num_processes=args.num_proc_bsearch, - cutoff_prob=args.cutoff_prob, - cutoff_top_n=args.cutoff_top_n, - ext_scoring_func=ext_scorer, ) - - result_transcripts = [res[0][1] for res in beam_search_results] - for target, result in zip(target_transcripts, result_transcripts): - err_sum[index] += error_rate_func(target, result) - err_ave[index] = err_sum[index] / num_ins - if index % 2 == 0: - sys.stdout.write('.') - sys.stdout.flush() - - # output on-line tuning result at the end of current batch - err_ave_min = min(err_ave) - min_index = err_ave.index(err_ave_min) - print("\nBatch %d [%d/?], current opt (alpha, beta) = (%s, %s), " - " min [%s] = %f" %(cur_batch, num_ins, - "%.3f" % params_grid[min_index][0], - "%.3f" % params_grid[min_index][1], - args.error_rate_type, err_ave_min)) - cur_batch += 1 - - # output WER/CER at every (alpha, beta) - print("\nFinal %s:\n" % args.error_rate_type) - for index in xrange(len(params_grid)): - print("(alpha, beta) = (%s, %s), [%s] = %f" - % ("%.3f" % params_grid[index][0], "%.3f" % params_grid[index][1], - args.error_rate_type, err_ave[index])) - - err_ave_min = min(err_ave) - min_index = err_ave.index(err_ave_min) - print("\nFinish tuning on %d batches, final opt (alpha, beta) = (%s, %s)" - % (args.num_batches, "%.3f" % params_grid[min_index][0], - "%.3f" % params_grid[min_index][1])) - - logger.info("finish tuning") - - -def main(): - print_arguments(args) - paddle.init(use_gpu=args.use_gpu, - rnn_use_batch=True, - trainer_count=args.trainer_count) - tune() - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/train.py b/deep_speech_2/train.py deleted file mode 100644 index 562fb46221be58065093eb9f9e6edd6ff97fc975..0000000000000000000000000000000000000000 --- a/deep_speech_2/train.py +++ /dev/null @@ -1,131 +0,0 @@ -"""Trainer for DeepSpeech2 model.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import argparse -import functools -import paddle.v2 as paddle -from model_utils.model import DeepSpeech2Model -from data_utils.data import DataGenerator -from utils.utility import add_arguments, print_arguments - -parser = argparse.ArgumentParser(description=__doc__) -add_arg = functools.partial(add_arguments, argparser=parser) -# yapf: disable -add_arg('batch_size', int, 256, "Minibatch size.") -add_arg('trainer_count', int, 8, "# of Trainers (CPUs or GPUs).") -add_arg('num_passes', int, 200, "# of training epochs.") -add_arg('num_proc_data', int, 16, "# of CPUs for data preprocessing.") -add_arg('num_conv_layers', int, 2, "# of convolution layers.") -add_arg('num_rnn_layers', int, 3, "# of recurrent layers.") -add_arg('rnn_layer_size', int, 2048, "# of recurrent cells per layer.") -add_arg('num_iter_print', int, 100, "Every # iterations for printing " - "train cost.") -add_arg('learning_rate', float, 5e-4, "Learning rate.") -add_arg('max_duration', float, 27.0, "Longest audio duration allowed.") -add_arg('min_duration', float, 0.0, "Shortest audio duration allowed.") -add_arg('test_off', bool, False, "Turn off testing.") -add_arg('use_sortagrad', bool, True, "Use SortaGrad or not.") -add_arg('use_gpu', bool, True, "Use GPU or not.") -add_arg('use_gru', bool, False, "Use GRUs instead of simple RNNs.") -add_arg('is_local', bool, True, "Use pserver or not.") -add_arg('share_rnn_weights',bool, True, "Share input-hidden weights across " - "bi-directional RNNs. Not for GRU.") -add_arg('train_manifest', str, - 'data/librispeech/manifest.train', - "Filepath of train manifest.") -add_arg('dev_manifest', str, - 'data/librispeech/manifest.dev-clean', - "Filepath of validation manifest.") -add_arg('mean_std_path', str, - 'data/librispeech/mean_std.npz', - "Filepath of normalizer's mean & std.") -add_arg('vocab_path', str, - 'data/librispeech/vocab.txt', - "Filepath of vocabulary.") -add_arg('init_model_path', str, - None, - "If None, the training starts from scratch, " - "otherwise, it resumes from the pre-trained model.") -add_arg('output_model_dir', str, - "./checkpoints/libri", - "Directory for saving checkpoints.") -add_arg('augment_conf_path',str, - 'conf/augmentation.config', - "Filepath of augmentation configuration file (json-format).") -add_arg('specgram_type', str, - 'linear', - "Audio feature type. Options: linear, mfcc.", - choices=['linear', 'mfcc']) -add_arg('shuffle_method', str, - 'batch_shuffle_clipped', - "Shuffle method.", - choices=['instance_shuffle', 'batch_shuffle', 'batch_shuffle_clipped']) -# yapf: disable -args = parser.parse_args() - - -def train(): - """DeepSpeech2 training.""" - train_generator = DataGenerator( - vocab_filepath=args.vocab_path, - mean_std_filepath=args.mean_std_path, - augmentation_config=open(args.augment_conf_path, 'r').read(), - max_duration=args.max_duration, - min_duration=args.min_duration, - specgram_type=args.specgram_type, - num_threads=args.num_proc_data, - num_conv_layers=args.num_conv_layers) - dev_generator = DataGenerator( - vocab_filepath=args.vocab_path, - mean_std_filepath=args.mean_std_path, - augmentation_config="{}", - specgram_type=args.specgram_type, - num_threads=args.num_proc_data, - num_conv_layers=args.num_conv_layers) - train_batch_reader = train_generator.batch_reader_creator( - manifest_path=args.train_manifest, - batch_size=args.batch_size, - min_batch_size=args.trainer_count, - sortagrad=args.use_sortagrad if args.init_model_path is None else False, - shuffle_method=args.shuffle_method) - dev_batch_reader = dev_generator.batch_reader_creator( - manifest_path=args.dev_manifest, - batch_size=args.batch_size, - min_batch_size=1, # must be 1, but will have errors. - sortagrad=False, - shuffle_method=None) - - ds2_model = DeepSpeech2Model( - vocab_size=train_generator.vocab_size, - num_conv_layers=args.num_conv_layers, - num_rnn_layers=args.num_rnn_layers, - rnn_layer_size=args.rnn_layer_size, - use_gru=args.use_gru, - pretrained_model_path=args.init_model_path, - share_rnn_weights=args.share_rnn_weights) - ds2_model.train( - train_batch_reader=train_batch_reader, - dev_batch_reader=dev_batch_reader, - feeding_dict=train_generator.feeding, - learning_rate=args.learning_rate, - gradient_clipping=400, - num_passes=args.num_passes, - num_iterations_print=args.num_iter_print, - output_model_dir=args.output_model_dir, - is_local=args.is_local, - test_off=args.test_off) - - -def main(): - print_arguments(args) - paddle.init(use_gpu=args.use_gpu, - rnn_use_batch=True, - trainer_count=args.trainer_count, - log_clipping=True) - train() - - -if __name__ == '__main__': - main() diff --git a/deep_speech_2/utils/__init__.py b/deep_speech_2/utils/__init__.py deleted file mode 100644 index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0000000000000000000000000000000000000000 diff --git a/deep_speech_2/utils/error_rate.py b/deep_speech_2/utils/error_rate.py deleted file mode 100644 index ea829f4703a90babd53e5408cfd30a427430de0d..0000000000000000000000000000000000000000 --- a/deep_speech_2/utils/error_rate.py +++ /dev/null @@ -1,154 +0,0 @@ -# -*- coding: utf-8 -*- -"""This module provides functions to calculate error rate in different level. -e.g. wer for word-level, cer for char-level. -""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import numpy as np - - -def _levenshtein_distance(ref, hyp): - """Levenshtein distance is a string metric for measuring the difference - between two sequences. Informally, the levenshtein disctance is defined as - the minimum number of single-character edits (substitutions, insertions or - deletions) required to change one word into the other. We can naturally - extend the edits to word level when calculate levenshtein disctance for - two sentences. - """ - m = len(ref) - n = len(hyp) - - # special case - if ref == hyp: - return 0 - if m == 0: - return n - if n == 0: - return m - - if m < n: - ref, hyp = hyp, ref - m, n = n, m - - # use O(min(m, n)) space - distance = np.zeros((2, n + 1), dtype=np.int32) - - # initialize distance matrix - for j in xrange(n + 1): - distance[0][j] = j - - # calculate levenshtein distance - for i in xrange(1, m + 1): - prev_row_idx = (i - 1) % 2 - cur_row_idx = i % 2 - distance[cur_row_idx][0] = i - for j in xrange(1, n + 1): - if ref[i - 1] == hyp[j - 1]: - distance[cur_row_idx][j] = distance[prev_row_idx][j - 1] - else: - s_num = distance[prev_row_idx][j - 1] + 1 - i_num = distance[cur_row_idx][j - 1] + 1 - d_num = distance[prev_row_idx][j] + 1 - distance[cur_row_idx][j] = min(s_num, i_num, d_num) - - return distance[m % 2][n] - - -def wer(reference, hypothesis, ignore_case=False, delimiter=' '): - """Calculate word error rate (WER). WER compares reference text and - hypothesis text in word-level. WER is defined as: - - .. math:: - WER = (Sw + Dw + Iw) / Nw - - where - - .. code-block:: text - - Sw is the number of words subsituted, - Dw is the number of words deleted, - Iw is the number of words inserted, - Nw is the number of words in the reference - - We can use levenshtein distance to calculate WER. Please draw an attention - that empty items will be removed when splitting sentences by delimiter. - - :param reference: The reference sentence. - :type reference: basestring - :param hypothesis: The hypothesis sentence. - :type hypothesis: basestring - :param ignore_case: Whether case-sensitive or not. - :type ignore_case: bool - :param delimiter: Delimiter of input sentences. - :type delimiter: char - :return: Word error rate. - :rtype: float - :raises ValueError: If the reference length is zero. - """ - if ignore_case == True: - reference = reference.lower() - hypothesis = hypothesis.lower() - - ref_words = filter(None, reference.split(delimiter)) - hyp_words = filter(None, hypothesis.split(delimiter)) - - if len(ref_words) == 0: - raise ValueError("Reference's word number should be greater than 0.") - - edit_distance = _levenshtein_distance(ref_words, hyp_words) - wer = float(edit_distance) / len(ref_words) - return wer - - -def cer(reference, hypothesis, ignore_case=False, remove_space=False): - """Calculate charactor error rate (CER). CER compares reference text and - hypothesis text in char-level. CER is defined as: - - .. math:: - CER = (Sc + Dc + Ic) / Nc - - where - - .. code-block:: text - - Sc is the number of characters substituted, - Dc is the number of characters deleted, - Ic is the number of characters inserted - Nc is the number of characters in the reference - - We can use levenshtein distance to calculate CER. Chinese input should be - encoded to unicode. Please draw an attention that the leading and tailing - space characters will be truncated and multiple consecutive space - characters in a sentence will be replaced by one space character. - - :param reference: The reference sentence. - :type reference: basestring - :param hypothesis: The hypothesis sentence. - :type hypothesis: basestring - :param ignore_case: Whether case-sensitive or not. - :type ignore_case: bool - :param remove_space: Whether remove internal space characters - :type remove_space: bool - :return: Character error rate. - :rtype: float - :raises ValueError: If the reference length is zero. - """ - if ignore_case == True: - reference = reference.lower() - hypothesis = hypothesis.lower() - - join_char = ' ' - if remove_space == True: - join_char = '' - - reference = join_char.join(filter(None, reference.split(' '))) - hypothesis = join_char.join(filter(None, hypothesis.split(' '))) - - if len(reference) == 0: - raise ValueError("Length of reference should be greater than 0.") - - edit_distance = _levenshtein_distance(reference, hypothesis) - cer = float(edit_distance) / len(reference) - return cer diff --git a/deep_speech_2/utils/tests/test_error_rate.py b/deep_speech_2/utils/tests/test_error_rate.py deleted file mode 100644 index d6bc7442e1f55bcea1f16234301785a884f2249a..0000000000000000000000000000000000000000 --- a/deep_speech_2/utils/tests/test_error_rate.py +++ /dev/null @@ -1,115 +0,0 @@ -# -*- coding: utf-8 -*- -"""Test error rate.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import unittest -from utils import error_rate - - -class TestParse(unittest.TestCase): - def test_wer_1(self): - ref = 'i UM the PHONE IS i LEFT THE portable PHONE UPSTAIRS last night' - hyp = 'i GOT IT TO the FULLEST i LOVE TO portable FROM OF STORES last '\ - 'night' - word_error_rate = error_rate.wer(ref, hyp) - self.assertTrue(abs(word_error_rate - 0.769230769231) < 1e-6) - - def test_wer_2(self): - ref = 'as any in england i would say said gamewell proudly that is '\ - 'in his day' - hyp = 'as any in england i would say said came well proudly that is '\ - 'in his day' - word_error_rate = error_rate.wer(ref, hyp) - self.assertTrue(abs(word_error_rate - 0.1333333) < 1e-6) - - def test_wer_3(self): - ref = 'the lieutenant governor lilburn w boggs afterward governor '\ - 'was a pronounced mormon hater and throughout the period of '\ - 'the troubles he manifested sympathy with the persecutors' - hyp = 'the lieutenant governor little bit how bags afterward '\ - 'governor was a pronounced warman hater and throughout the '\ - 'period of th troubles he manifests sympathy with the '\ - 'persecutors' - word_error_rate = error_rate.wer(ref, hyp) - self.assertTrue(abs(word_error_rate - 0.2692307692) < 1e-6) - - def test_wer_4(self): - ref = 'the wood flamed up splendidly under the large brewing copper '\ - 'and it sighed so deeply' - hyp = 'the wood flame do splendidly under the large brewing copper '\ - 'and its side so deeply' - word_error_rate = error_rate.wer(ref, hyp) - self.assertTrue(abs(word_error_rate - 0.2666666667) < 1e-6) - - def test_wer_5(self): - ref = 'all the morning they trudged up the mountain path and at noon '\ - 'unc and ojo sat on a fallen tree trunk and ate the last of '\ - 'the bread which the old munchkin had placed in his pocket' - hyp = 'all the morning they trudged up the mountain path and at noon '\ - 'unc in ojo sat on a fallen tree trunk and ate the last of '\ - 'the bread which the old munchkin had placed in his pocket' - word_error_rate = error_rate.wer(ref, hyp) - self.assertTrue(abs(word_error_rate - 0.027027027) < 1e-6) - - def test_wer_6(self): - ref = 'i UM the PHONE IS i LEFT THE portable PHONE UPSTAIRS last night' - word_error_rate = error_rate.wer(ref, ref) - self.assertEqual(word_error_rate, 0.0) - - def test_wer_7(self): - ref = ' ' - hyp = 'Hypothesis sentence' - with self.assertRaises(ValueError): - word_error_rate = error_rate.wer(ref, hyp) - - def test_cer_1(self): - ref = 'werewolf' - hyp = 'weae wolf' - char_error_rate = error_rate.cer(ref, hyp) - self.assertTrue(abs(char_error_rate - 0.25) < 1e-6) - - def test_cer_2(self): - ref = 'werewolf' - hyp = 'weae wolf' - char_error_rate = error_rate.cer(ref, hyp, remove_space=True) - self.assertTrue(abs(char_error_rate - 0.125) < 1e-6) - - def test_cer_3(self): - ref = 'were wolf' - hyp = 'were wolf' - char_error_rate = error_rate.cer(ref, hyp) - self.assertTrue(abs(char_error_rate - 0.0) < 1e-6) - - def test_cer_4(self): - ref = 'werewolf' - char_error_rate = error_rate.cer(ref, ref) - self.assertEqual(char_error_rate, 0.0) - - def test_cer_5(self): - ref = u'我是中国人' - hyp = u'我是 美洲人' - char_error_rate = error_rate.cer(ref, hyp) - self.assertTrue(abs(char_error_rate - 0.6) < 1e-6) - - def test_cer_6(self): - ref = u'我 是 中 国 人' - hyp = u'我 是 美 洲 人' - char_error_rate = error_rate.cer(ref, hyp, remove_space=True) - self.assertTrue(abs(char_error_rate - 0.4) < 1e-6) - - def test_cer_7(self): - ref = u'我是中国人' - char_error_rate = error_rate.cer(ref, ref) - self.assertFalse(char_error_rate, 0.0) - - def test_cer_8(self): - ref = '' - hyp = 'Hypothesis' - with self.assertRaises(ValueError): - char_error_rate = error_rate.cer(ref, hyp) - - -if __name__ == '__main__': - unittest.main() diff --git a/deep_speech_2/utils/utility.py b/deep_speech_2/utils/utility.py deleted file mode 100644 index 2e489ade6f28fdce5c6b60b47bc919a55549f046..0000000000000000000000000000000000000000 --- a/deep_speech_2/utils/utility.py +++ /dev/null @@ -1,47 +0,0 @@ -"""Contains common utility functions.""" -from __future__ import absolute_import -from __future__ import division -from __future__ import print_function - -import distutils.util - - -def print_arguments(args): - """Print argparse's arguments. - - Usage: - - .. code-block:: python - - parser = argparse.ArgumentParser() - parser.add_argument("name", default="Jonh", type=str, help="User name.") - args = parser.parse_args() - print_arguments(args) - - :param args: Input argparse.Namespace for printing. - :type args: argparse.Namespace - """ - print("----------- Configuration Arguments -----------") - for arg, value in sorted(vars(args).iteritems()): - print("%s: %s" % (arg, value)) - print("------------------------------------------------") - - -def add_arguments(argname, type, default, help, argparser, **kwargs): - """Add argparse's argument. - - Usage: - - .. code-block:: python - - parser = argparse.ArgumentParser() - add_argument("name", str, "Jonh", "User name.", parser) - args = parser.parse_args() - """ - type = distutils.util.strtobool if type == bool else type - argparser.add_argument( - "--" + argname, - default=default, - type=type, - help=help + ' Default: %(default)s.', - **kwargs) diff --git a/deep_speech_2/utils/utility.sh b/deep_speech_2/utils/utility.sh deleted file mode 100644 index baae04743410a2d9e9827b738abd7893fe2458e9..0000000000000000000000000000000000000000 --- a/deep_speech_2/utils/utility.sh +++ /dev/null @@ -1,23 +0,0 @@ -download() { - URL=$1 - MD5=$2 - TARGET=$3 - - if [ -e $TARGET ]; then - md5_result=`md5sum $TARGET | awk -F[' '] '{print $1}'` - if [ $MD5 == $md5_result ]; then - echo "$TARGET already exists, download skipped." - return 0 - fi - fi - - wget -c $URL -O "$TARGET" - if [ $? -ne 0 ]; then - return 1 - fi - - md5_result=`md5sum $TARGET | awk -F[' '] '{print $1}'` - if [ ! $MD5 == $md5_result ]; then - return 1 - fi -} diff --git a/dssm/README.cn.md b/dssm/README.cn.md index e1bd3cab89a2f76752c946b5a54cc250def9441b..4a80c87673e958c958c82a2cd5dfb7bf0dbaa075 100644 --- a/dssm/README.cn.md +++ b/dssm/README.cn.md @@ -131,7 +131,7 @@ def create_cnn(self, emb, prefix=''): conv_3 = create_conv(3, self.dnn_dims[1], "cnn") conv_4 = create_conv(4, self.dnn_dims[1], "cnn") - return conv_3, conv_4 + return paddle.layer.concat(input=[conv_3, conv_4]) ``` CNN 接受词向量序列，通过卷积和池化操作捕捉到原始句子的关键信息，最终输出一个语义向量（可以认为是句子向量）。 @@ -216,54 +216,54 @@ Pairwise Rank复用上面的DNN结构，同一个source对两个target求相似 ### 回归的数据格式 ``` # 3 fields each line: -# - source's word ids -# - target's word ids +# - source word list +# - target word list # - target - \t \t + \t \t ``` 比如： ``` -3 6 10 \t 6 8 33 \t 0.7 -6 0 \t 6 9 330 \t 0.03 +苹果 六 袋 苹果 6s 0.1 +新手 汽车 驾驶 驾校 培训 0.9 ``` ### 分类的数据格式 ``` # 3 fields each line: -# - source's word ids -# - target's word ids +# - source word list +# - target word list # - target - \t \t