diff --git a/PaddleCV/image_classification/build_model.py b/PaddleCV/image_classification/build_model.py index 003111ab86206729ae20d56fca304943f16038da..6b374b17a5a66c3da766fa35f725ee187aaae541 100644 --- a/PaddleCV/image_classification/build_model.py +++ b/PaddleCV/image_classification/build_model.py @@ -15,7 +15,7 @@ import paddle import paddle.fluid as fluid import utils.utility as utility -AMP_MODEL_LIST = ["ResNet50", "SE_ResNet50_vd"] +AMP_MODEL_LIST = ["ResNet50", "SE_ResNet50_vd", "ResNet200_vd"] def _calc_label_smoothing_loss(softmax_out, label, class_dim, epsilon): diff --git a/PaddleCV/image_classification/models/resnet_vd.py b/PaddleCV/image_classification/models/resnet_vd.py index bb04e2f6e73d8f67fae42d6e6666044db4cf3c34..97544931ad6f2c8d901ef5a0fa35fe5dc763bdc7 100644 --- a/PaddleCV/image_classification/models/resnet_vd.py +++ b/PaddleCV/image_classification/models/resnet_vd.py @@ -23,7 +23,8 @@ import paddle.fluid as fluid from paddle.fluid.param_attr import ParamAttr __all__ = [ - "ResNet", "ResNet18_vd", "ResNet34_vd", "ResNet50_vd", "ResNet101_vd", "ResNet152_vd", "ResNet200_vd" + "ResNet", "ResNet18_vd", "ResNet34_vd", "ResNet50_vd", "ResNet101_vd", + "ResNet152_vd", "ResNet200_vd" ] @@ -32,7 +33,7 @@ class ResNet(): self.layers = layers self.is_3x3 = is_3x3 - def net(self, input, class_dim=1000): + def net(self, input, class_dim=1000, data_format="NCHW"): is_3x3 = self.is_3x3 layers = self.layers supported_layers = [18, 34, 50, 101, 152, 200] @@ -40,7 +41,7 @@ class ResNet(): "supported layers are {} but input layer is {}".format(supported_layers, layers) if layers == 18: - depth = [2, 2, 2, 2] + depth = [2, 2, 2, 2] elif layers == 34 or layers == 50: depth = [3, 4, 6, 3] elif layers == 101: @@ -56,7 +57,8 @@ class ResNet(): num_filters=64, filter_size=7, stride=2, - act='relu') + act='relu', + data_format=data_format) else: conv = self.conv_bn_layer( input=input, @@ -64,29 +66,33 @@ class ResNet(): filter_size=3, stride=2, act='relu', - name='conv1_1') + name='conv1_1', + data_format=data_format) conv = self.conv_bn_layer( input=conv, num_filters=32, filter_size=3, stride=1, act='relu', - name='conv1_2') + name='conv1_2', + data_format=data_format) conv = self.conv_bn_layer( input=conv, num_filters=64, filter_size=3, stride=1, act='relu', - name='conv1_3') + name='conv1_3', + data_format=data_format) conv = fluid.layers.pool2d( input=conv, pool_size=3, pool_stride=2, pool_padding=1, - pool_type='max') - + pool_type='max', + data_format=data_format) + if layers >= 50: for block in range(len(depth)): for i in range(depth[block]): @@ -101,22 +107,29 @@ class ResNet(): input=conv, num_filters=num_filters[block], stride=2 if i == 0 and block != 0 else 1, - if_first=block==i==0, - name=conv_name) + if_first=block == i == 0, + name=conv_name, + data_format=data_format) else: for block in range(len(depth)): for i in range(depth[block]): - conv_name="res"+str(block+2)+chr(97+i) + conv_name = "res" + str(block + 2) + chr(97 + i) conv = self.basic_block( input=conv, num_filters=num_filters[block], stride=2 if i == 0 and block != 0 else 1, - if_first=block==i==0, - name=conv_name) + if_first=block == i == 0, + name=conv_name, + data_format=data_format) pool = fluid.layers.pool2d( - input=conv, pool_type='avg', global_pooling=True) - stdv = 1.0 / math.sqrt(pool.shape[1] * 1.0) + input=conv, + pool_type='avg', + global_pooling=True, + data_format=data_format) + pool_channel = pool.shape[1] if data_format == "NCHW" else pool.shape[ + -1] + stdv = 1.0 / math.sqrt(pool_channel * 1.0) out = fluid.layers.fc( input=pool, @@ -133,7 +146,8 @@ class ResNet(): stride=1, groups=1, act=None, - name=None): + name=None, + data_format="NCHW"): conv = fluid.layers.conv2d( input=input, num_filters=num_filters, @@ -143,7 +157,8 @@ class ResNet(): groups=groups, act=None, param_attr=ParamAttr(name=name + "_weights"), - bias_attr=False) + bias_attr=False, + data_format=data_format) if name == "conv1": bn_name = "bn_" + name else: @@ -154,7 +169,8 @@ class ResNet(): param_attr=ParamAttr(name=bn_name + '_scale'), bias_attr=ParamAttr(bn_name + '_offset'), moving_mean_name=bn_name + '_mean', - moving_variance_name=bn_name + '_variance') + moving_variance_name=bn_name + '_variance', + data_layout=data_format) def conv_bn_layer_new(self, input, @@ -163,14 +179,16 @@ class ResNet(): stride=1, groups=1, act=None, - name=None): + name=None, + data_format="NCHW"): pool = fluid.layers.pool2d( input=input, pool_size=2, pool_stride=2, pool_padding=0, pool_type='avg', - ceil_mode=True) + ceil_mode=True, + data_format=data_format) conv = fluid.layers.conv2d( input=pool, @@ -181,7 +199,8 @@ class ResNet(): groups=groups, act=None, param_attr=ParamAttr(name=name + "_weights"), - bias_attr=False) + bias_attr=False, + data_format=data_format) if name == "conv1": bn_name = "bn_" + name else: @@ -192,81 +211,114 @@ class ResNet(): param_attr=ParamAttr(name=bn_name + '_scale'), bias_attr=ParamAttr(bn_name + '_offset'), moving_mean_name=bn_name + '_mean', - moving_variance_name=bn_name + '_variance') + moving_variance_name=bn_name + '_variance', + data_layout=data_format) - def shortcut(self, input, ch_out, stride, name, if_first=False): - ch_in = input.shape[1] + def shortcut(self, + input, + ch_out, + stride, + name, + if_first=False, + data_format="NCHW"): + ch_in = input.shape[1] if data_format == "NCHW" else input.shape[-1] if ch_in != ch_out or stride != 1: if if_first: - return self.conv_bn_layer(input, ch_out, 1, stride, name=name) + return self.conv_bn_layer( + input, + ch_out, + 1, + stride, + name=name, + data_format=data_format) else: - return self.conv_bn_layer_new(input, ch_out, 1, stride, name=name) + return self.conv_bn_layer_new( + input, + ch_out, + 1, + stride, + name=name, + data_format=data_format) elif if_first: - return self.conv_bn_layer(input, ch_out, 1, stride, name=name) + return self.conv_bn_layer( + input, ch_out, 1, stride, name=name, data_format=data_format) else: return input - - def bottleneck_block(self, input, num_filters, stride, name, if_first): + def bottleneck_block(self, + input, + num_filters, + stride, + name, + if_first, + data_format="NCHW"): conv0 = self.conv_bn_layer( input=input, num_filters=num_filters, filter_size=1, act='relu', - name=name + "_branch2a") + name=name + "_branch2a", + data_format=data_format) conv1 = self.conv_bn_layer( input=conv0, num_filters=num_filters, filter_size=3, stride=stride, act='relu', - name=name + "_branch2b") + name=name + "_branch2b", + data_format=data_format) conv2 = self.conv_bn_layer( input=conv1, num_filters=num_filters * 4, filter_size=1, act=None, - name=name + "_branch2c") + name=name + "_branch2c", + data_format=data_format) short = self.shortcut( input, num_filters * 4, stride, if_first=if_first, - name=name + "_branch1") + name=name + "_branch1", + data_format=data_format) return fluid.layers.elementwise_add(x=short, y=conv2, act='relu') - - - def basic_block(self, input, num_filters, stride, name, if_first): + + def basic_block(self, input, num_filters, stride, name, if_first, + data_format): conv0 = self.conv_bn_layer( - input=input, - num_filters=num_filters, - filter_size=3, - act='relu', + input=input, + num_filters=num_filters, + filter_size=3, + act='relu', stride=stride, - name=name+"_branch2a") + name=name + "_branch2a", + data_format=data_format) conv1 = self.conv_bn_layer( - input=conv0, - num_filters=num_filters, - filter_size=3, - act=None, - name=name+"_branch2b") + input=conv0, + num_filters=num_filters, + filter_size=3, + act=None, + name=name + "_branch2b", + data_format=data_format) short = self.shortcut( - input, - num_filters, - stride, - if_first=if_first, - name=name + "_branch1") + input, + num_filters, + stride, + if_first=if_first, + name=name + "_branch1", + data_format=data_format) return fluid.layers.elementwise_add(x=short, y=conv1, act='relu') + def ResNet18_vd(): - model=ResNet(layers=18, is_3x3=True) + model = ResNet(layers=18, is_3x3=True) return model def ResNet34_vd(): - model=ResNet(layers=34, is_3x3=True) + model = ResNet(layers=34, is_3x3=True) return model diff --git a/PaddleCV/image_classification/scripts/train/ResNet200_vd_fp16.sh b/PaddleCV/image_classification/scripts/train/ResNet200_vd_fp16.sh new file mode 100755 index 0000000000000000000000000000000000000000..4631cb357c8a35850d62629288cbdc2746883b7f --- /dev/null +++ b/PaddleCV/image_classification/scripts/train/ResNet200_vd_fp16.sh @@ -0,0 +1,49 @@ +#!/bin/bash -ex + +#Training details +export FLAGS_conv_workspace_size_limit=4000 #MB +export FLAGS_cudnn_exhaustive_search=1 +export FLAGS_cudnn_batchnorm_spatial_persistent=1 + +DATA_DIR="Your image dataset path, e.g. ./data/ILSVRC2012/" +DATA_FORMAT="NHWC" +USE_AMP=true #whether to use amp +USE_DALI=true +USE_ADDTO=true + +if ${USE_ADDTO} ;then + export FLAGS_max_inplace_grad_add=8 +fi +if ${USE_DALI}; then + export FLAGS_fraction_of_gpu_memory_to_use=0.8 +fi + +python train.py \ + --model=ResNet200_vd \ + --data_dir=${DATA_DIR} \ + --batch_size=64 \ + --num_epochs=200 \ + --total_images=1281167 \ + --image_shape 4 224 224 \ + --class_dim=1000 \ + --print_step=10 \ + --model_save_dir=output/ \ + --lr_strategy=cosine_decay \ + --use_amp=${USE_AMP} \ + --scale_loss=128.0 \ + --use_dynamic_loss_scaling=true \ + --data_format=${DATA_FORMAT} \ + --fuse_elewise_add_act_ops=true \ + --fuse_bn_act_ops=true \ + --fuse_bn_add_act_ops=true \ + --enable_addto=${USE_ADDTO} \ + --validate=true \ + --is_profiler=false \ + --profiler_path=profile/ \ + --reader_thread=10 \ + --reader_buf_size=4000 \ + --use_dali=${USE_DALI} \ + --lr=0.1 \ + --l2_decay=1e-4 \ + --use_label_smoothing=True \ + --label_smoothing_epsilon=0.1 diff --git a/PaddleNLP/README_en.md b/PaddleNLP/README_en.md index 5575d59691086e5594da7d8f824bf2db068c6b29..15a9d3343d10763390a38716a111c5125bb537fc 100644 --- a/PaddleNLP/README_en.md +++ b/PaddleNLP/README_en.md @@ -95,13 +95,16 @@ For more pretrained model selection, please refer to [PretrainedModels](./paddle - [Models API](./docs/models.md) + + + ## Tutorials Please refer to our official AI Studio account for more interactive tutorials: [PaddleNLP on AI Studio](https://aistudio.baidu.com/aistudio/personalcenter/thirdview/574995) -* [What's Seq2Vec?](https://aistudio.baidu.com/aistudio/projectdetail/1294333) shows how to use LSTM to do sentiment analysis. +* [What's Seq2Vec?](https://aistudio.baidu.com/aistudio/projectdetail/1283423) shows how to use LSTM to do sentiment analysis. -* [Sentiment Analysis with ERNIE](https://aistudio.baidu.com/aistudio/projectdetail/1283423) shows how to exploit the pretrained ERNIE to make sentiment analysis better. +* [Sentiment Analysis with ERNIE](https://aistudio.baidu.com/aistudio/projectdetail/1294333) shows how to exploit the pretrained ERNIE to make sentiment analysis better. * [Waybill Information Extraction with BiGRU-CRF Model](https://aistudio.baidu.com/aistudio/projectdetail/1317771) shows how to make use of bigru and crf to do information extraction. diff --git a/PaddleNLP/benchmark/bert/README.md b/PaddleNLP/benchmark/bert/README.md index b5e92687163c14a2ff40c8584143c0809a448cab..0948c3e677fa4c2c64572b2d6e22a922a508556c 100644 --- a/PaddleNLP/benchmark/bert/README.md +++ b/PaddleNLP/benchmark/bert/README.md @@ -1,6 +1,6 @@ # BERT Benchmark with Fleet API BERT - Bidirectional Encoder Representations from Transformers [论文链接](https://arxiv.org/abs/1810.04805) -PaddlePaddle实现了BERT的预训练模型(Pre-training)和下游任务(Fine-tunning)。在预训练任务上提供单机版本和多机版本,同时提供混合精度接口来进行加速,可以任务需要进行选择。 +PaddlePaddle实现了BERT的预训练模型(Pre-training)和下游任务(Fine-tunning)。 ## 数据集 ### Pre-training数据集 @@ -10,7 +10,8 @@ PaddlePaddle实现了BERT的预训练模型(Pre-training)和下游任务(Fin ## Pre-training任务训练 ### 环境变量设置 1. paddlenlp的安装 -pip install paddlenlp==2.0.0a2 -i https://pypi.org/simple +pip install paddlenlp==2.0.0b0 -i https://pypi.org/simple + 2. 设置预训练的数据地址环境变量 ```shell export DATA_DIR=${HOME}/bert_data/wikicorpus_en @@ -54,26 +55,6 @@ python ./run_pretrain_single.py \ --max_steps 1000000 ``` -### 训练速度对比 -进行速度对比的模型是bert-based模型,主要对比的方式是单机单机和多机多卡(4机32卡)下面进行速度对比,所有的GPU测试配置都是基于 Tesla V100-SXM2-16GB,下面的配置如下: -- InfiniBand 100 Gb/sec (4X EDR), Mellanox Technologies MT27700 Family -- 48 CPU(s), Intel(R) Xeon(R) Gold 5118 CPU @ 2.30GHz -- Memory 500G -- Ubuntu 16.04.4 LTS (GNU/Linux 4.4.0-116-generic x86_64) -- CUDA Version: 10.2, Driver API Version: 10.2, Driver Version: 440.33.01 -- cuDNN Version: 7.6 -- PaddlePaddle version: paddlepadle-gpu >= 2.0.0rc1 -- PaddleNLP version: paddlenlp >= 2.0.0a2 - -速度统计方式是统计每秒预训练模型能处理的样本数量,其中 -- batch_size=64 -- max_seq_length=128 - -下面是具体速度对比情况: -| node num | node num | gpu num/node | gpu num | batch_size/gpu |Throughput | Speedup | -|----------| -------- | -------------| ------- | -------- | ----------| ------- | - - ## Fine-tuning任务训练 在完成 BERT 模型的预训练后,即可利用预训练参数在特定的 NLP 任务上做 Fine-tuning。以下利用开源的预训练模型,示例如何进行分类任务的 Fine-tuning。 diff --git a/PaddleNLP/examples/README.md b/PaddleNLP/examples/README.md index 6f3be9805cd5db26d5b5cb14460e95163fe7ed41..af8d47ee88c006cf88cca4fa0b57b554b5d0e9be 100644 --- a/PaddleNLP/examples/README.md +++ b/PaddleNLP/examples/README.md @@ -7,16 +7,16 @@ | 任务类型 | 目录 | 简介 | | ----------------------------------| ------------------------------------------------------------ | ------------------------------------------------------------ | -| 中文词法分析 | [LAC(Lexical Analysis of Chinese)](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/lexical_analysis) | 百度自主研发中文特色模型词法分析任务,集成了中文分词、词性标注和命名实体识别任务。输入是一个字符串,而输出是句子中的词边界和词性、实体类别。 | +| 中文词法分析 | [LAC(Lexical Analysis of Chinese)](./lexical_analysis) | 百度自主研发中文特色模型词法分析任务,集成了中文分词、词性标注和命名实体识别任务。输入是一个字符串,而输出是句子中的词边界和词性、实体类别。 | | 预训练词向量 | [WordEmbedding](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/word_embedding) | 提供了丰富的中文预训练词向量,通过简单配置即可使用词向量来进行热启训练,能支持较多的中文场景下的训练任务的热启训练,加快训练收敛速度。| ### 核心技术模型 | 任务类型 | 目录 | 简介 | | -------------------------------- | ------------------------------------------------------------ | ------------------------------------------------------------ | -| ERNIE-GEN文本生成 | [ERNIE-GEN(An Enhanced Multi-Flow Pre-training and Fine-tuning Framework for Natural Language Generation)](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/text_generation/ernie-gen) |ERNIE-GEN是百度发布的生成式预训练模型,是一种Multi-Flow结构的预训练和微调框架。ERNIE-GEN利用更少的参数量和数据,在摘要生成、问题生成、对话和生成式问答4个任务共5个数据集上取得了SOTA效果 | -| BERT 预训练&GLUE下游任务 | [BERT(Bidirectional Encoder Representation from Transformers)](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/bert) | BERT模型作为目前最为火热语义表示预训练模型,PaddleNLP提供了简洁功效的实现方式,同时易用性方面通过简单参数切换即可实现不同的BERT模型。 | -| Electra 预训练&GLUE下游任务 | [Electra(Pre-training Text Encoders as Discriminators Rather Than Generator)](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/electra) |ELECTRA模型新一种模型预训练的框架,采用generator和discriminator的结合方式,相对于BERT来说能提升计算效率,同时缓解BERT训练和预测不一致的问题。| +| ERNIE-GEN文本生成 | [ERNIE-GEN(An Enhanced Multi-Flow Pre-training and Fine-tuning Framework for Natural Language Generation)](./text_generation/ernie-gen) |ERNIE-GEN是百度发布的生成式预训练模型,是一种Multi-Flow结构的预训练和微调框架。ERNIE-GEN利用更少的参数量和数据,在摘要生成、问题生成、对话和生成式问答4个任务共5个数据集上取得了SOTA效果 | +| BERT 预训练&GLUE下游任务 | [BERT(Bidirectional Encoder Representation from Transformers)](./language_model/bert) | BERT模型作为目前最为火热语义表示预训练模型,PaddleNLP提供了简洁功效的实现方式,同时易用性方面通过简单参数切换即可实现不同的BERT模型。 | +| Electra 预训练&GLUE下游任务 | [Electra(Efficiently Learning an Encoder that Classifies Token Replacements Accurately)](./language_model/electra) |ELECTRA 创新性地引入GAN的思想对BERT预训练过程进行了改进,在和BERT具有相同的模型参数、预训练计算量一样的情况下,ELECTRA GLUE得分明显好。同时相比GPT、ELMo,在GLUE得分略好时,ELECTRA预训练模型只需要很少的参数和计算量。| ### 核心应用模型 @@ -25,20 +25,20 @@ | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [Seq2Seq](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/machine_translation/seq2seq) | 使用编码器-解码器(Encoder-Decoder)结构, 同时使用了Attention机制来加强Decoder和Encoder之间的信息交互,Seq2Seq 广泛应用于机器翻译,自动对话机器人,文档摘要自动生成,图片描述自动生成等任务中。| -| [Transformer](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/machine_translation/transformer) |基于PaddlePaddle框架的Transformer结构搭建的机器翻译模型,Transformer 计算并行度高,能解决学习长程依赖问题。并且模型框架集成了训练,验证,预测任务,功能完备,效果突出。| +| [Seq2Seq](./machine_translation/seq2seq) | 使用编码器-解码器(Encoder-Decoder)结构, 同时使用了Attention机制来加强Decoder和Encoder之间的信息交互,Seq2Seq 广泛应用于机器翻译,自动对话机器人,文档摘要自动生成,图片描述自动生成等任务中。| +| [Transformer](./machine_translation/transformer) |基于PaddlePaddle框架的Transformer结构搭建的机器翻译模型,Transformer 计算并行度高,能解决学习长程依赖问题。并且模型框架集成了训练,验证,预测任务,功能完备,效果突出。| #### 命名实体识别 (Named Entity Recognition) 命名实体识别(Named Entity Recognition,NER)是NLP中一项非常基础的任务。NER是信息提取、问答系统、句法分析、机器翻译等众多NLP任务的重要基础工具。命名实体识别的准确度,决定了下游任务的效果,是NLP中非常重要的一个基础问题。 在NER任务提供了两种解决方案,一类LSTM/GRU + CRF(Conditional Random Field),RNN类的模型来抽取底层文本的信息,而CRF(条件随机场)模型来学习底层Token之间的联系;另外一类是通过预训练模型,例如ERNIE,BERT模型,直接来预测Token的标签信息。 -因为该类模型较为抽象,提供了一份快递单信息抽取的训练脚本给大家使用,具体的任务是通过两类的模型来抽取快递单的核心信息,例如地址,姓名,手机号码,具体的[快递单任务链接](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/named_entity_recognition/express_ner)。 +因为该类模型较为抽象,提供了一份快递单信息抽取的训练脚本给大家使用,具体的任务是通过两类的模型来抽取快递单的核心信息,例如地址,姓名,手机号码,具体的[快递单任务链接](./named_entity_recognition/express_ner)。 下面是具体的模型信息。 | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [BiGRU+CRF](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/named_entity_recognition/express_ner) |传统的序列标注模型,通过双向GRU模型能抽取文本序列的信息和联系,通过CRF模型来学习文本Token之间的联系,本模型集成PaddleNLP自己开发的CRF模型,模型结构清晰易懂。 | -| [ERNIE/BERT Fine-tuning](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/named_entity_recognition) |通过预训练模型提供的强大的语义信息和ERNIE/BERT类模型的Self-Attention机制来覆盖Token之间的联系,直接通过BERT/ERNIE的序列分类模型来预测文本每个token的标签信息,模型结构简单,效果优异。| +| [BiGRU-CRF](./named_entity_recognition/express_ner) |传统的序列标注模型,通过双向GRU模型能抽取文本序列的信息和联系,通过CRF模型来学习文本Token之间的联系,本模型集成PaddleNLP自己开发的CRF模型,模型结构清晰易懂。 | +| [ERNIE/BERT Fine-tuning](./named_entity_recognition) |通过预训练模型提供的强大的语义信息和ERNIE/BERT类模型的Self-Attention机制来覆盖Token之间的联系,直接通过BERT/ERNIE的序列分类模型来预测文本每个token的标签信息,模型结构简单,效果优异。| #### 文本分类 (Text Classification) @@ -46,8 +46,8 @@ | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [RNN/GRU/LSTM](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/text_classification/rnn) | 面向通用场景的文本分类模型,网络结构接入常见的RNN类模型,例如LSTM,GRU,RNN。整体模型结构集成在百度的自研的Senta文本情感分类模型上,效果突出,用法简易。| -| [ERNIE/BERT Fine-tuning](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/text_classification/pretrained_models) |基于预训练后模型的文本分类的模型,多达11种的预训练模型可供使用,其中有较多中文预训练模型,预训练模型切换简单,情感分析任务上效果突出。| +| [RNN/GRU/LSTM](./text_classification/rnn) | 面向通用场景的文本分类模型,网络结构接入常见的RNN类模型,例如LSTM,GRU,RNN。整体模型结构集成在百度的自研的Senta文本情感分类模型上,效果突出,用法简易。| +| [ERNIE/BERT Fine-tuning](./text_classification/pretrained_models) |基于预训练后模型的文本分类的模型,多达11种的预训练模型可供使用,其中有较多中文预训练模型,预训练模型切换简单,情感分析任务上效果突出。| #### 文本生成 (Text Generation) @@ -55,7 +55,7 @@ | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [ERNIE-GEN(An Enhanced Multi-Flow Pre-training and Fine-tuning Framework for Natural Language Generation)](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/text_generation/ernie-gen) |ERNIE-GEN是百度发布的生成式预训练模型,通过Global-Attention的方式解决训练和预测曝光偏差的问题,同时使用Multi-Flow Attention机制来分别进行Global和Context信息的交互,同时通过片段生成的方式来增加语义相关性。| +| [ERNIE-GEN(An Enhanced Multi-Flow Pre-training and Fine-tuning Framework for Natural Language Generation)](./text_generation/ernie-gen) |ERNIE-GEN是百度发布的生成式预训练模型,通过Global-Attention的方式解决训练和预测曝光偏差的问题,同时使用Multi-Flow Attention机制来分别进行Global和Context信息的交互,同时通过片段生成的方式来增加语义相关性。| @@ -65,8 +65,8 @@ | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [SimNet](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/text_matching/simnet)|PaddleNLP提供的SimNet模型已经纳入了PaddleNLP的官方API中,用户直接调用API即完成一个SimNet模型的组网,在模型层面提供了Bow/CNN/LSTM/GRU常用信息抽取方式, 灵活高,使用方便。| -| [SentenceTransformer](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/text_matching/sentence_transformers)|直接调用简易的预训练模型接口接口完成对Sentence的语义表示,同时提供了较多的中文预训练模型,可以根据任务的来选择相关参数。| +| [SimNet](./text_matching/simnet)|PaddleNLP提供的SimNet模型已经纳入了PaddleNLP的官方API中,用户直接调用API即完成一个SimNet模型的组网,在模型层面提供了Bow/CNN/LSTM/GRU常用信息抽取方式, 灵活高,使用方便。| +| [SentenceTransformer](./text_matching/sentence_transformers)|直接调用简易的预训练模型接口接口完成对Sentence的语义表示,同时提供了较多的中文预训练模型,可以根据任务的来选择相关参数。| #### 语言模型 (Language Model) @@ -74,8 +74,8 @@ | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [RNNLM](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/language_model/rnnlm) |序列任务常用的rnn网络,实现了一个两层的LSTM网络,然后LSTM的结果去预测下一个词出现的概率。是基于RNN的常规的语言模型。| -| [ELMo](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/language_model/elmo) |ElMo是一个双向的LSTM语言模型,由一个前向和一个后向语言模型构成,目标函数就是取这两个方向语言模型的最大似然。ELMo主要是解决了传统的WordEmbedding的向量表示单一的问题,ELMo通过结合上下文来增强语义表示。| +| [RNNLM](./language_model/rnnlm) |序列任务常用的rnn网络,实现了一个两层的LSTM网络,然后LSTM的结果去预测下一个词出现的概率。是基于RNN的常规的语言模型。| +| [ELMo](./language_model/elmo) |ElMo是一个双向的LSTM语言模型,由一个前向和一个后向语言模型构成,目标函数就是取这两个方向语言模型的最大似然。ELMo主要是解决了传统的WordEmbedding的向量表示单一的问题,ELMo通过结合上下文来增强语义表示。| #### 文本图学习 (Text Graph) 在很多工业应用中,往往出现一种特殊的图:Text Graph。顾名思义,图的节点属性由文本构成,而边的构建提供了结构信息。如搜索场景下的Text Graph,节点可由搜索词、网页标题、网页正文来表达,用户反馈和超链信息则可构成边关系。百度图学习PGL((Paddle Graph Learning)团队提出ERNIESage(ERNIE SAmple aggreGatE)模型同时建模文本语义与图结构信息,有效提升Text Graph的应用效果。图学习是深度学习领域目前的研究热点,如果想对图学习有更多的了解,可以访问[PGL Github链接](https://github.com/PaddlePaddle/PGL/)。 @@ -83,14 +83,14 @@ ERNIESage模型的具体信息如下。 | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [ERNIESage(ERNIE SAmple aggreGatE)](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/text_graph/erniesage)|通过Graph(图)来来构建自身节点和邻居节点的连接关系,将自身节点和邻居节点的关系构建成一个关联样本输入到ERNIE中,ERNIE作为聚合函数(Aggregators)来表征自身节点和邻居节点的语义关系,最终强化图中节点的语义表示。在TextGraph的任务上ERNIESage的效果非常优秀。| +| [ERNIESage(ERNIE SAmple aggreGatE)](./text_graph/erniesage)|通过Graph(图)来来构建自身节点和邻居节点的连接关系,将自身节点和邻居节点的关系构建成一个关联样本输入到ERNIE中,ERNIE作为聚合函数(Aggregators)来表征自身节点和邻居节点的语义关系,最终强化图中节点的语义表示。在TextGraph的任务上ERNIESage的效果非常优秀。| #### 阅读理解(Machine Reading Comprehension) 机器阅读理解是近期自然语言处理领域的研究热点之一,也是人工智能在处理和理解人类语言进程中的一个长期目标。得益于深度学习技术和大规模标注数据集的发展,用端到端的神经网络来解决阅读理解任务取得了长足的进步。下面是具体的模型信息。 | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [BERT Fine-tuning](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/machine_reading_comprehension/) |通过ERNIE/BERT等预训练模型的强大的语义表示能力,设置在阅读理解上面的下游任务,该模块主要是提供了多个数据集来验证BERT模型在阅读理解上的效果,数据集主要是包括了SQuAD,DuReader,DuReader-robust,DuReader-yesno。同时提供了和相关阅读理解相关的Metric(指标),用户可以简易的调用这些API,快速验证模型效果。| +| [BERT Fine-tuning](./machine_reading_comprehension/) |通过ERNIE/BERT等预训练模型的强大的语义表示能力,设置在阅读理解上面的下游任务,该模块主要是提供了多个数据集来验证BERT模型在阅读理解上的效果,数据集主要是包括了SQuAD,DuReader,DuReader-robust,DuReader-yesno。同时提供了和相关阅读理解相关的Metric(指标),用户可以简易的调用这些API,快速验证模型效果。| #### 对话系统(Dialogue System) @@ -98,7 +98,8 @@ ERNIESage模型的具体信息如下。 | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [BERT-DGU](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/dialogue/dgu) |通过ERNIE/BERT等预训练模型的强大的语义表示能力,抽取对话中的文本语义信息,通过对文本分类等操作就可以完成对话中的诸多任务,例如意图识别,行文识别,状态跟踪等。| +| [DGU](./dialogue/dgu) |通过ERNIE/BERT等预训练模型的强大的语义表示能力,抽取对话中的文本语义信息,通过对文本分类等操作就可以完成对话中的诸多任务,例如意图识别,行文识别,状态跟踪等。| +| [PLATO-2](./dialogue/plato-2) | 百度自研领先的开放域对话预训练模型。[PLATO-2: Towards Building an Open-Domain Chatbot via Curriculum Learning](https://arxiv.org/abs/2006.16779) | #### 时间序列预测(Time Series) 时间序列是指按照时间先后顺序排列而成的序列,例如每日发电量、每小时营业额等组成的序列。通过分析时间序列中的发展过程、方向和趋势,我们可以预测下一段时间可能出现的情况。为了更好让大家了解时间序列预测任务,提供了基于19年新冠疫情预测的任务示例,有兴趣的话可以进行研究学习。 @@ -107,4 +108,4 @@ ERNIESage模型的具体信息如下。 | 模型 | 简介 | | ------------------------------------------------------------ | ------------------------------------------------------------ | -| [TCN(Temporal convolutional network)](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/examples/time_series)|TCN模型基于卷积的时间序列模型,通过因果卷积(Causal Convolution)和空洞卷积(Dilated Convolution) 特定的组合方式解决卷积不适合时间序列任务的问题,TCN具备并行度高,内存低等诸多优点,在某些时间序列任务上效果已经超过传统的RNN模型。| +| [TCN(Temporal convolutional network)](./time_series)|TCN模型基于卷积的时间序列模型,通过因果卷积(Causal Convolution)和空洞卷积(Dilated Convolution) 特定的组合方式解决卷积不适合时间序列任务的问题,TCN具备并行度高,内存低等诸多优点,在某些时间序列任务上效果已经超过传统的RNN模型。| diff --git a/PaddleNLP/examples/dialogue/dgu/README.md b/PaddleNLP/examples/dialogue/dgu/README.md index 8f7d1f0c3f5910e12dc3547381cdadc702884138..6f52035e327dda42db426e993501510bb46fd6e7 100644 --- a/PaddleNLP/examples/dialogue/dgu/README.md +++ b/PaddleNLP/examples/dialogue/dgu/README.md @@ -39,17 +39,17 @@ DGU模型中的6个任务,分别采用不同的评估指标在test集上进行 * PaddlePaddle 安装 - 本项目依赖于 PaddlePaddle 2.0 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 + 本项目依赖于 PaddlePaddle 2.0rc1 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 * PaddleNLP 安装 ```shell - pip install paddlenlp + pip install paddlenlp>=2.0.0b ``` * 环境依赖 - Python的版本要求 3.6+,其它环境请参考 PaddlePaddle [安装说明](https://www.paddlepaddle.org.cn/install/quick/zh/2.0rc-linux-docker) 部分的内容 + Python的版本要求 3.6+ ### 代码结构说明 diff --git a/PaddleNLP/examples/dialogue/plato-2/README.md b/PaddleNLP/examples/dialogue/plato-2/README.md index a7b0f68a0cae399045a1f9dc21bc46cf982d16b7..132f00e609ed7d37cfb87ef1bbd2e03b6f772c8b 100644 --- a/PaddleNLP/examples/dialogue/plato-2/README.md +++ b/PaddleNLP/examples/dialogue/plato-2/README.md @@ -18,7 +18,7 @@ PLATO-2的训练过程及其他细节详见 [Knover](https://github.com/PaddlePa * PaddlePaddle 安装 - 本项目依赖于 PaddlePaddle 2.0 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 + 本项目依赖于 PaddlePaddle 2.0rc1 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 * PaddleNLP 安装 @@ -28,13 +28,13 @@ PLATO-2的训练过程及其他细节详见 [Knover](https://github.com/PaddlePa * 环境依赖 - Python的版本要求 3.6+ + Python的版本要求 3.6+ - 本项目依赖sentencepiece和termcolor,请在运行本项目之前进行安装 + 本项目依赖sentencepiece和termcolor,请在运行本项目之前进行安装 - ```shell - pip install sentencepiece termcolor - ``` + ```shell + pip install sentencepiece termcolor + ``` ### 代码结构说明 diff --git a/PaddleNLP/examples/electra/README.md b/PaddleNLP/examples/language_model/electra/README.md similarity index 100% rename from PaddleNLP/examples/electra/README.md rename to PaddleNLP/examples/language_model/electra/README.md diff --git a/PaddleNLP/examples/electra/run_glue.py b/PaddleNLP/examples/language_model/electra/run_glue.py similarity index 100% rename from PaddleNLP/examples/electra/run_glue.py rename to PaddleNLP/examples/language_model/electra/run_glue.py diff --git a/PaddleNLP/examples/electra/run_pretrain.py b/PaddleNLP/examples/language_model/electra/run_pretrain.py similarity index 100% rename from PaddleNLP/examples/electra/run_pretrain.py rename to PaddleNLP/examples/language_model/electra/run_pretrain.py diff --git a/PaddleNLP/examples/language_model/elmo/README.md b/PaddleNLP/examples/language_model/elmo/README.md index 4e9b81fbf5d5f49160904b489b04d6cdcd10ea63..d126ea2a4e5d28d75be6d8ff7d84bc92a20b145c 100644 --- a/PaddleNLP/examples/language_model/elmo/README.md +++ b/PaddleNLP/examples/language_model/elmo/README.md @@ -18,15 +18,17 @@ ELMo(Embeddings from Language Models)是重要的通用语义表示模型之一 * PaddlePaddle 安装 - 本项目依赖于 PaddlePaddle 2.0 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 + 本项目依赖于 PaddlePaddle 2.0rc1 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 * 环境依赖 - Python的版本要求 3.6+,并安装sklearn和gensim。其它环境请参考 PaddlePaddle [安装说明](https://www.paddlepaddle.org.cn/documentation/docs/zh/1.5/beginners_guide/install/index_cn.html) 部分的内容 + Python的版本要求 3.6+ -```shell -pip install sklearn gensim -``` + 本项目依赖sklearn和gensim,请在运行本项目之前进行安装 + + ```shell + pip install sklearn gensim + ``` ### 代码结构说明 diff --git a/PaddleNLP/examples/named_entity_recognition/README.md b/PaddleNLP/examples/named_entity_recognition/README.md index 537d9ba93e7ac6fcfc721089c573c20f96043e89..923c9d35889e2e951d9d6bc25a0fa4c16ee4f4b9 100644 --- a/PaddleNLP/examples/named_entity_recognition/README.md +++ b/PaddleNLP/examples/named_entity_recognition/README.md @@ -1,66 +1,7 @@ -# Name Entity Recognition +# 命名实体识别 -## 快递单信息抽取 +命名实体识别(Named Entity Recognition,简称NER),又称作“专名识别”,是指识别文本中具有特定意义的实体,主要包括人名、地名、机构名、专有名词等,它是信息提取、问答系统、句法分析、机器翻译、面向Semantic Web的元数据标注等应用领域的重要基础工具,在自然语言处理技术走向实用化的过程中占有重要地位。在本例中,我们将介绍使用PaddleNLP运行开源数据集MSRA_NER,同时我们还将介绍一个有趣的应用例子——快递单信息抽取。 -## Part1. Bi-LSTM+CRF NER +* [MSRA_NER](msra_ner/) -## Part2. BERT NER - -### 序列标注任务 - -以 MSRA 任务为例,启动 Fine-tuning 的方式如下(`paddlenlp` 要已经安装或能在 `PYTHONPATH` 中找到): - -```shell -export CUDA_VISIBLE_DEVICES=0 - -python -u ./run_msra_ner.py \ - --model_name_or_path bert-base-multilingual-uncased \ - --max_seq_length 128 \ - --batch_size 32 \ - --learning_rate 2e-5 \ - --num_train_epochs 3 \ - --logging_steps 1 \ - --save_steps 500 \ - --output_dir ./tmp/msra_ner/ \ - --n_gpu 1 -``` - -其中参数释义如下: -- `model_name_or_path` 指示了某种特定配置的模型,对应有其预训练模型和预训练时使用的 tokenizer。若模型相关内容保存在本地,这里也可以提供相应目录地址。 -- `max_seq_length` 表示最大句子长度,超过该长度将被截断。 -- `batch_size` 表示每次迭代**每张卡**上的样本数目。 -- `learning_rate` 表示基础学习率大小,将于learning rate scheduler产生的值相乘作为当前学习率。 -- `num_train_epochs` 表示训练轮数。 -- `logging_steps` 表示日志打印间隔。 -- `save_steps` 表示模型保存及评估间隔。 -- `output_dir` 表示模型保存路径。 -- `n_gpu` 表示使用的 GPU 卡数。若希望使用多卡训练,将其设置为指定数目即可;若为0,则使用CPU。 - -训练过程将按照 `logging_steps` 和 `save_steps` 的设置打印如下日志: - -``` -global step 996, epoch: 1, batch: 344, loss: 0.038471, speed: 4.72 step/s -global step 997, epoch: 1, batch: 345, loss: 0.032820, speed: 4.82 step/s -global step 998, epoch: 1, batch: 346, loss: 0.008144, speed: 4.69 step/s -global step 999, epoch: 1, batch: 347, loss: 0.031425, speed: 4.36 step/s -global step 1000, epoch: 1, batch: 348, loss: 0.073151, speed: 4.59 step/s -eval loss: 0.019874, precision: 0.991670, recall: 0.991930, f1: 0.991800 -``` - -使用以上命令进行单卡 Fine-tuning ,在验证集上有如下结果: - Metric | Result | -------------------------------|-------------| -precision | 0.992903 | -recall | 0.991823 | -f1 | 0.992363 | - -# TODO: 写成教程 -参考run_bert_crf.py,进一步使用CRF - Metric | Result | -------------------------------|-------------| -precision | 0.992266 | -recall | 0.993056 | -f1 | 0.992661 | - - -## Part3. BERT+LSTM-CRF NER +* [快递单信息抽取](express_ner/) diff --git a/PaddleNLP/examples/named_entity_recognition/express_ner/README.md b/PaddleNLP/examples/named_entity_recognition/express_ner/README.md new file mode 100644 index 0000000000000000000000000000000000000000..f0b61033e3f4028296ffad067fca5257e5a657fd --- /dev/null +++ b/PaddleNLP/examples/named_entity_recognition/express_ner/README.md @@ -0,0 +1,60 @@ +# 快递单信息抽取 + +## 1. 简介 + +本项目将演示如何从用户提供的快递单中,抽取姓名、电话、省、市、区、详细地址等内容,形成结构化信息。辅助物流行业从业者进行有效信息的提取,从而降低客户填单的成本。 + +## 2. 快速开始 + +### 2.1 环境配置 + +- Python >= 3.6 + +- paddlepaddle >= 2.0.0rc1,安装方式请参考 [快速安装](https://www.paddlepaddle.org.cn/install/quick)。 + +- paddlenlp >= 2.0.0b, 安装方式:`pip install paddlenlp>=2.0.0b` + + +### 2.2 数据准备 + +数据集已经保存在data目录中,示例如下 + +``` +16620200077宣荣嗣甘肃省白银市会宁县河畔镇十字街金海超市西行50米 T-BT-IT-IT-IT-IT-IT-IT-IT-IT-IT-IP-BP-IP-IA1-BA1-IA1-IA2-BA2-IA2-IA3-BA3-IA3-IA4-BA4-IA4-IA4-IA4-IA4-IA4-IA4-IA4-IA4-IA4-IA4-IA4-IA4-IA4-I +13552664307姜骏炜云南省德宏傣族景颇族自治州盈江县平原镇蜜回路下段 T-BT-IT-IT-IT-IT-IT-IT-IT-IT-IT-IP-BP-IP-IA1-BA1-IA1-IA2-BA2-IA2-IA2-IA2-IA2-IA2-IA2-IA2-IA2-IA3-BA3-IA3-IA4-BA4-IA4-IA4-IA4-IA4-IA4-IA4-I +``` +数据集中以特殊字符"\t"分隔文本、标签,以特殊字符"\002"分隔每个字。标签的定义如下: + +| 标签 | 定义 | 标签 | 定义 | +| -------- | -------- |-------- | -------- | +| P-B | 姓名起始位置 | P-I | 姓名中间位置或结束位置 | +| T-B | 电话起始位置 | T-I | 电话中间位置或结束位置 | +| A1-B | 省份起始位置 | A1-I | 省份中间位置或结束位置 | +| A2-B | 城市起始位置 | A2-I | 城市中间位置或结束位置 | +| A3-B | 县区起始位置 | A3-I | 县区中间位置或结束位置 | +| A4-B | 详细地址起始位置 | A4-I | 详细地址中间位置或结束位置 | +| O | 无关字符 | | | + +注意每个标签的结果只有 B、I、O 三种,这种标签的定义方式叫做 BIO 体系。其中 B 表示一个标签类别的开头,比如 P-B 指的是姓名的开头;相应的,I 表示一个标签的延续。 + +### 2.3 启动训练 + +本项目提供了两种模型结构,一种是BiGRU + CRF结构,另一种是ERNIE + FC结构,前者显存占用小,后者能够在较小的迭代次数中收敛。 + +#### 2.3.1 启动BiGRU + CRF训练 + +```bash +export CUDA_VISIBLE_DEVICES=0 # 只支持单卡训练 +python run_bigru_crf.py +``` + +详细介绍请参考教程:[基于Bi-GRU+CRF的快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) + +#### 2.3.2 启动ERNIE + FC训练 + +```bash +export CUDA_VISIBLE_DEVICES=0 # 只支持单卡训练 +python run_ernie.py +``` + +详细介绍请参考教程:[使用PaddleNLP预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) diff --git a/PaddleNLP/examples/named_entity_recognition/msra_ner/README.md b/PaddleNLP/examples/named_entity_recognition/msra_ner/README.md new file mode 100644 index 0000000000000000000000000000000000000000..8a3e3ee6628c83c7252ed1af8ee2e71314499fff --- /dev/null +++ b/PaddleNLP/examples/named_entity_recognition/msra_ner/README.md @@ -0,0 +1,73 @@ +# 使用PaddleNLP运行MSRA-NER + +## 1. 简介 + +MSRA-NER 数据集由微软亚研院发布,其目标是识别文本中具有特定意义的实体,主要包括人名、地名、机构名等。示例如下: + +``` +海钓比赛地点在厦门与金门之间的海域。 OOOOOOOB-LOCI-LOCOB-LOCI-LOCOOOOOO +这座依山傍水的博物馆由国内一流的设计师主持设计,整个建筑群精美而恢宏。 OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO +``` + +数据集中以特殊字符"\t"分隔文本、标签,以特殊字符"\002"分隔每个字。 + +## 2. 快速开始 + +### 2.1 环境配置 + +- Python >= 3.6 + +- paddlepaddle >= 2.0.0rc1,安装方式请参考 [快速安装](https://www.paddlepaddle.org.cn/install/quick)。 + +- paddlenlp >= 2.0.0b, 安装方式:`pip install paddlenlp>=2.0.0b` + +### 2.2 启动MSRA-NER任务 + +```shell +export CUDA_VISIBLE_DEVICES=0 + +python -u ./run_msra_ner.py \ + --model_name_or_path bert-base-multilingual-uncased \ + --max_seq_length 128 \ + --batch_size 32 \ + --learning_rate 2e-5 \ + --num_train_epochs 3 \ + --logging_steps 1 \ + --save_steps 500 \ + --output_dir ./tmp/msra_ner/ \ + --n_gpu 1 +``` + +其中参数释义如下: +- `model_name_or_path` 指示了某种特定配置的模型,对应有其预训练模型和预训练时使用的 tokenizer。若模型相关内容保存在本地,这里也可以提供相应目录地址。 +- `max_seq_length` 表示最大句子长度,超过该长度将被截断。 +- `batch_size` 表示每次迭代**每张卡**上的样本数目。 +- `learning_rate` 表示基础学习率大小,将于learning rate scheduler产生的值相乘作为当前学习率。 +- `num_train_epochs` 表示训练轮数。 +- `logging_steps` 表示日志打印间隔。 +- `save_steps` 表示模型保存及评估间隔。 +- `output_dir` 表示模型保存路径。 +- `n_gpu` 表示使用的 GPU 卡数。若希望使用多卡训练,将其设置为指定数目即可;若为0,则使用CPU。 + +训练过程将按照 `logging_steps` 和 `save_steps` 的设置打印如下日志: + +``` +global step 996, epoch: 1, batch: 344, loss: 0.038471, speed: 4.72 step/s +global step 997, epoch: 1, batch: 345, loss: 0.032820, speed: 4.82 step/s +global step 998, epoch: 1, batch: 346, loss: 0.008144, speed: 4.69 step/s +global step 999, epoch: 1, batch: 347, loss: 0.031425, speed: 4.36 step/s +global step 1000, epoch: 1, batch: 348, loss: 0.073151, speed: 4.59 step/s +eval loss: 0.019874, precision: 0.991670, recall: 0.991930, f1: 0.991800 +``` + +使用以上命令进行单卡 Fine-tuning ,在验证集上有如下结果: + Metric | Result | +------------------------------|-------------| +precision | 0.992903 | +recall | 0.991823 | +f1 | 0.992363 | + +## 参考 + +[Microsoft Research Asia Chinese Word-Segmentation Data Set](https://www.microsoft.com/en-us/download/details.aspx?id=52531) +[The third international Chinese language processing bakeoff: Word segmentation and named entity recognition](https://faculty.washington.edu/levow/papers/sighan06.pdf) diff --git a/PaddleNLP/examples/named_entity_recognition/run_msra_ner.py b/PaddleNLP/examples/named_entity_recognition/msra_ner/run_msra_ner.py similarity index 100% rename from PaddleNLP/examples/named_entity_recognition/run_msra_ner.py rename to PaddleNLP/examples/named_entity_recognition/msra_ner/run_msra_ner.py diff --git a/PaddleNLP/examples/text_classification/README.md b/PaddleNLP/examples/text_classification/README.md index 0eb6c93d4ed829e2525a64ccb51a0e5c25649e9e..f52a6a2bec4454fbb5aa2f159359ffbbe8ba5d94 100644 --- a/PaddleNLP/examples/text_classification/README.md +++ b/PaddleNLP/examples/text_classification/README.md @@ -4,7 +4,7 @@ ## Conventional RNNs Models -[Recurrent Neural Networks](./rnn) 展示了如何使用RNN、LSTM、GRU等网络完成文本分类任务。 +[Recurrent Neural Networks](./rnn) 展示了如何使用传统序列模型RNN、LSTM、GRU等网络完成文本分类任务。 ## Pretrained Model (PTMs) @@ -12,10 +12,18 @@ ## 线上体验教程 -* [paddlenlp.seq2vec是什么? 瞧瞧它怎么完成情感分析教程](https://aistudio.baidu.com/aistudio/projectdetail/1294333)展示了使用序列模型LSTM完成情感分析任务。 +- [使用seq2vec模块进行句子情感分类](https://aistudio.baidu.com/aistudio/projectdetail/1283423) -* [使用PaddleNLP语义预训练模型ERNIE优化情感分析教程](https://aistudio.baidu.com/aistudio/projectdetail/1283423)展示了使用ERNIE优化情感分析任务。 +- [如何将预训练模型Fine-tune下游任务](https://aistudio.baidu.com/aistudio/projectdetail/1294333) -* [基于Bi-GRU+CRF的快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) +- [使用Bi-GRU+CRF完成快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) -* [使用PaddleNLP预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) +- [使用预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) + +- [使用Seq2Seq模型完成自动对联模型](https://aistudio.baidu.com/aistudio/projectdetail/1321118) + +- [使用预训练模型ERNIE-GEN实现智能写诗](https://aistudio.baidu.com/aistudio/projectdetail/1339888) + +- [使用TCN网络完成新冠疫情病例数预测](https://aistudio.baidu.com/aistudio/projectdetail/1290873) + +更多教程参见[PaddleNLP on AI Studio](https://aistudio.baidu.com/aistudio/personalcenter/thirdview/574995)。 diff --git a/PaddleNLP/examples/text_classification/pretrained_models/README.md b/PaddleNLP/examples/text_classification/pretrained_models/README.md index 658112cbc927cb5aa2ddddd12a201c3915304ad1..47efa9ce19d28b436a22b422dfb2602917c8ef66 100644 --- a/PaddleNLP/examples/text_classification/pretrained_models/README.md +++ b/PaddleNLP/examples/text_classification/pretrained_models/README.md @@ -1,17 +1,34 @@ # 使用预训练模型Fine-tune完成中文文本分类任务 -随着深度学习的发展,模型参数的数量飞速增长。为了训练这些参数,需要更大的数据集来避免过拟合。然而,对于大部分NLP任务来说,构建大规模的标注数据集非常困难(成本过高),特别是对于句法和语义相关的任务。相比之下,大规模的未标注语料库的构建则相对容易。为了利用这些数据,我们可以先从其中学习到一个好的表示,再将这些表示应用到其他任务中。最近的研究表明,基于大规模未标注语料库的预训练模型(Pretrained Models, PTM) 在NLP任务上取得了很好的表现。 + +在2017年之前,工业界和学术界对NLP文本处理依赖于序列模型[Recurrent Neural Network (RNN)](../rnn). + +

+
+

+ + +[paddlenlp.seq2vec是什么? 瞧瞧它怎么完成情感分析](https://aistudio.baidu.com/aistudio/projectdetail/1283423)教程介绍了如何使用`paddlenlp.seq2vec`表征文本语义。 + +近年来随着深度学习的发展,模型参数的数量飞速增长。为了训练这些参数,需要更大的数据集来避免过拟合。然而,对于大部分NLP任务来说,构建大规模的标注数据集非常困难(成本过高),特别是对于句法和语义相关的任务。相比之下,大规模的未标注语料库的构建则相对容易。为了利用这些数据,我们可以先从其中学习到一个好的表示,再将这些表示应用到其他任务中。最近的研究表明,基于大规模未标注语料库的预训练模型(Pretrained Models, PTM) 在NLP任务上取得了很好的表现。 近年来,大量的研究表明基于大型语料库的预训练模型(Pretrained Models, PTM)可以学习通用的语言表示,有利于下游NLP任务,同时能够避免从零开始训练模型。随着计算能力的发展,深度模型的出现(即 Transformer)和训练技巧的增强使得 PTM 不断发展,由浅变深。 -本示例展示了以BERT([Bidirectional Encoder Representations from Transformers](https://arxiv.org/abs/1810.04805))代表的预训练模型如何Finetune完成中文文本分类任务。 + +

+
+

+ +本图片来自于:https://github.com/thunlp/PLMpapers + +本示例展示了以ERNIE([Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223))代表的预训练模型如何Finetune完成中文文本分类任务。 ## 模型简介 本项目针对中文文本分类问题,开源了一系列模型,供用户可配置地使用: + BERT([Bidirectional Encoder Representations from Transformers](https://arxiv.org/abs/1810.04805))中文模型,简写`bert-base-chinese`, 其由12层Transformer网络组成。 -+ ERNIE([Enhanced Representation through Knowledge Integration](https://arxiv.org/pdf/1904.09223)),支持ERNIE 1.0中文模型(简写`ernie-1.0`)和ERNIE Tiny中文模型(简写`ernie-tiny`)。 ++ ERNIE([Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223)),支持ERNIE 1.0中文模型(简写`ernie-1.0`)和ERNIE Tiny中文模型(简写`ernie-tiny`)。 其中`ernie`由12层Transformer网络组成,`ernie-tiny`由3层Transformer网络组成。 + RoBERTa([A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692)),支持24层Transformer网络的`roberta-wwm-ext-large`和12层Transformer网络的`roberta-wwm-ext`。 @@ -29,21 +46,14 @@ ## 快速开始 -### 安装说明 - -* PaddlePaddle 安装 +### 环境依赖 - 本项目依赖于 PaddlePaddle 2.0 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 +- python >= 3.6 +- paddlepaddle >= 2.0.0-rc1 -* PaddleNLP 安装 - - ```shell - pip install paddlenlp - ``` - -* 环境依赖 - - Python的版本要求 3.6+,其它环境请参考 PaddlePaddle [安装说明](https://www.paddlepaddle.org.cn/documentation/docs/zh/1.5/beginners_guide/install/index_cn.html) 部分的内容 +``` +pip install paddlenlp==2.0.0b +``` ### 代码结构说明 @@ -128,10 +138,18 @@ Data: 作为老的四星酒店,房间依然很整洁,相当不错。机场 ## 线上体验教程 -* [paddlenlp.seq2vec是什么? 瞧瞧它怎么完成情感分析教程](https://aistudio.baidu.com/aistudio/projectdetail/1294333)展示了使用序列模型LSTM完成情感分析任务。 +- [使用seq2vec模块进行句子情感分类](https://aistudio.baidu.com/aistudio/projectdetail/1283423) + +- [如何将预训练模型Fine-tune下游任务](https://aistudio.baidu.com/aistudio/projectdetail/1294333) + +- [使用Bi-GRU+CRF完成快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) + +- [使用预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) + +- [使用Seq2Seq模型完成自动对联模型](https://aistudio.baidu.com/aistudio/projectdetail/1321118) -* [使用PaddleNLP语义预训练模型ERNIE优化情感分析教程](https://aistudio.baidu.com/aistudio/projectdetail/1283423)展示了使用ERNIE优化情感分析任务。 +- [使用预训练模型ERNIE-GEN实现智能写诗](https://aistudio.baidu.com/aistudio/projectdetail/1339888) -* [基于Bi-GRU+CRF的快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) +- [使用TCN网络完成新冠疫情病例数预测](https://aistudio.baidu.com/aistudio/projectdetail/1290873) -* [使用PaddleNLP预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) +更多教程参见[PaddleNLP on AI Studio](https://aistudio.baidu.com/aistudio/personalcenter/thirdview/574995)。 diff --git a/PaddleNLP/examples/text_classification/rnn/README.md b/PaddleNLP/examples/text_classification/rnn/README.md index d41980d6d804219c5b2a5f4c2a89b25afbd64549..32e2a4707e0ef784b432115878c3ce0908a86e20 100644 --- a/PaddleNLP/examples/text_classification/rnn/README.md +++ b/PaddleNLP/examples/text_classification/rnn/README.md @@ -2,19 +2,73 @@ 文本分类是NLP应用最广的任务之一,可以被应用到多个领域中,包括但不仅限于:情感分析、垃圾邮件识别、商品评价分类... -一般通过将文本表示成向量后接入分类器,完成文本分类。 +情感分析是一个自然语言处理中老生常谈的任务。情感分析的目的是为了找出说话者/作者在某些话题上,或者针对一个文本两极的观点的态度。这个态度或许是他或她的个人判断或是评估,也许是他当时的情感状态(就是说,作者在做出这个言论时的情绪状态),或是作者有意向的情感交流(就是作者想要读者所体验的情绪)。其可以用于数据挖掘、Web 挖掘、文本挖掘和信息检索方面得到了广泛的研究。可通过 [AI开放平台-情感倾向分析](http://ai.baidu.com/tech/nlp_apply/sentiment_classify) 线上体验。 -如何用向量表征文本,使得向量携带语义信息,是我们关心的重点。 +

+
+

本项目开源了一系列模型用于进行文本建模,用户可通过参数配置灵活使用。效果上,我们基于开源情感倾向分类数据集ChnSentiCorp对多个模型进行评测。 -情感倾向分析(Sentiment Classification)是一类常见的文本分类任务。其针对带有主观描述的中文文本,可自动判断该文本的情感极性类别并给出相应的置信度。情感类型分为积极、消极。情感倾向分析能够帮助企业理解用户消费习惯、分析热点话题和危机舆情监控,为企业提供有利的决策支持。可通过 [AI开放平台-情感倾向分析](http://ai.baidu.com/tech/nlp_apply/sentiment_classify) 线上体验。 +## paddlenlp.seq2vec + +情感分析任务中关键技术是如何将文本表示成一个**携带语义的文本向量**。随着深度学习技术的快速发展,目前常用的文本表示技术有LSTM,GRU,RNN等方法。 +PaddleNLP提供了一系列的文本表示技术,如`seq2vec`模块。 + +[`paddlenlp.seq2vec`](../../../paddlenlp/seq2vec) 模块作用为将输入的序列文本表征成一个语义向量。 + +

+
+

+ ## 模型简介 -本项目通过调用[Seq2Vec](../../../paddlenlp/seq2vec/)中内置的模型进行序列建模,完成句子的向量表示。包含最简单的词袋模型和一系列经典的RNN类模型。 +本项目通过调用[seq2vec](../../../paddlenlp/seq2vec/)中内置的模型进行序列建模,完成句子的向量表示。包含最简单的词袋模型和一系列经典的RNN类模型。 + +`seq2vec`模块 + +* 功能是将序列Embedding Tensor(shape是(batch_size, num_token, emb_dim) )转化成文本语义表征Enocded Texts Tensor(shape 是(batch_sie,encoding_size)) +* 提供了`BoWEncoder`,`CNNEncoder`,`GRUEncoder`,`LSTMEncoder`,`RNNEncoder`等模型 + - `BoWEncoder` 是将输入序列Embedding Tensor在num_token维度上叠加,得到文本语义表征Enocded Texts Tensor。 + - `CNNEncoder` 是将输入序列Embedding Tensor进行卷积操作,在对卷积结果进行max_pooling,得到文本语义表征Enocded Texts Tensor。 + - `GRUEncoder` 是对输入序列Embedding Tensor进行GRU运算,在运算结果上进行pooling或者取最后一个step的隐表示,得到文本语义表征Enocded Texts Tensor。 + - `LSTMEncoder` 是对输入序列Embedding Tensor进行LSTM运算,在运算结果上进行pooling或者取最后一个step的隐表示,得到文本语义表征Enocded Texts Tensor。 + - `RNNEncoder` 是对输入序列Embedding Tensor进行RNN运算,在运算结果上进行pooling或者取最后一个step的隐表示,得到文本语义表征Enocded Texts Tensor。 + + +`seq2vec`提供了许多语义表征方法,那么这些方法在什么时候更加适合呢? + +* `BoWEncoder`采用Bag of Word Embedding方法,其特点是简单。但其缺点是没有考虑文本的语境,所以对文本语义的表征不足以表意。 + +* `CNNEncoder`采用卷积操作,提取局部特征,其特点是可以共享权重。但其缺点同样只考虑了局部语义,上下文信息没有充分利用。 + +

+
+

+ +* `RNNEnocder`采用RNN方法,在计算下一个token语义信息时,利用上一个token语义信息作为其输入。但其缺点容易产生梯度消失和梯度爆炸。 + +

+
+

+ +* `LSTMEnocder`采用LSTM方法,LSTM是RNN的一种变种。为了学到长期依赖关系,LSTM 中引入了门控机制来控制信息的累计速度, + 包括有选择地加入新的信息,并有选择地遗忘之前累计的信息。 + +

+
+

+ +* `GRUEncoder`采用GRU方法,GRU也是RNN的一种变种。一个LSTM单元有四个输入 ,因而参数是RNN的四倍,带来的结果是训练速度慢。 + GRU对LSTM进行了简化,在不影响效果的前提下加快了训练速度。 + +

+
+

+ | 模型 | 模型介绍 | | ------------------------------------------------ | ------------------------------------------------------------ | @@ -38,25 +92,31 @@ | Bi-LSTM Attention | 0.8992 | 0.8856 | | TextCNN | 0.9102 | 0.9107 | -## 快速开始 -### 安装说明 +

+
+

-* PaddlePaddle 安装 - 本项目依赖于 PaddlePaddle 2.0 及以上版本,请参考 [安装指南](http://www.paddlepaddle.org/#quick-start) 进行安装 +关于CNN、LSTM、GRU、RNN等更多信息参考: -* PaddleNLP 安装 +* https://canvas.stanford.edu/files/1090785/download +* https://colah.github.io/posts/2015-08-Understanding-LSTMs/ +* https://arxiv.org/abs/1412.3555 +* https://arxiv.org/pdf/1506.00019 +* https://arxiv.org/abs/1404.2188 - ```shell - pip install paddlenlp - ``` -* 环境依赖 +## 快速开始 - 本项目依赖于jieba分词,请在运行本项目之前,安装jieba,如`pip install -U jieba` +### 环境依赖 - Python的版本要求 3.6+,其它环境请参考 PaddlePaddle [安装说明](https://www.paddlepaddle.org.cn/install/quick/zh/2.0rc-linux-docker) 部分的内容 +- python >= 3.6 +- paddlepaddle >= 2.0.0-rc1 + +``` +pip install paddlenlp==2.0.0b +``` ### 代码结构说明 @@ -164,10 +224,18 @@ Data: 作为老的四星酒店,房间依然很整洁,相当不错。机场 ## 线上体验教程 -* [paddlenlp.seq2vec是什么? 瞧瞧它怎么完成情感分析教程](https://aistudio.baidu.com/aistudio/projectdetail/1294333)展示了使用序列模型LSTM完成情感分析任务。 +- [使用seq2vec模块进行句子情感分类](https://aistudio.baidu.com/aistudio/projectdetail/1283423) + +- [如何将预训练模型Fine-tune下游任务](https://aistudio.baidu.com/aistudio/projectdetail/1294333) + +- [使用Bi-GRU+CRF完成快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) + +- [使用预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) + +- [使用Seq2Seq模型完成自动对联模型](https://aistudio.baidu.com/aistudio/projectdetail/1321118) -* [使用PaddleNLP语义预训练模型ERNIE优化情感分析教程](https://aistudio.baidu.com/aistudio/projectdetail/1283423)展示了使用ERNIE优化情感分析任务。 +- [使用预训练模型ERNIE-GEN实现智能写诗](https://aistudio.baidu.com/aistudio/projectdetail/1339888) -* [基于Bi-GRU+CRF的快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) +- [使用TCN网络完成新冠疫情病例数预测](https://aistudio.baidu.com/aistudio/projectdetail/1290873) -* [使用PaddleNLP预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) +更多教程参见[PaddleNLP on AI Studio](https://aistudio.baidu.com/aistudio/personalcenter/thirdview/574995)。 diff --git a/PaddleNLP/examples/text_generation/ernie-gen/README.md b/PaddleNLP/examples/text_generation/ernie-gen/README.md index 861ca92163d4f05ab9150145b77d00aabf2d87c3..f95c487b528ba5c5a0f2d3f51894f3d3642389c3 100644 --- a/PaddleNLP/examples/text_generation/ernie-gen/README.md +++ b/PaddleNLP/examples/text_generation/ernie-gen/README.md @@ -124,3 +124,9 @@ python -u ./predict.py \ year={2020} } ``` + +## 线上教程体验 + +我们为诗歌文本生成提供了线上教程,欢迎体验: + +* [使用PaddleNLP预训练模型ERNIE-GEN生成诗歌](https://aistudio.baidu.com/aistudio/projectdetail/1339888) diff --git a/PaddleNLP/examples/text_matching/README.md b/PaddleNLP/examples/text_matching/README.md index 7b84d6b1959949e9214dc156de5b27384f22496c..f51e74afdf43ddcebc0831daff08f89876369786 100644 --- a/PaddleNLP/examples/text_matching/README.md +++ b/PaddleNLP/examples/text_matching/README.md @@ -24,3 +24,21 @@ ## Sentence Transformers [Sentence Transformers](./sentence_transformers) 展示了如何使用以ERNIE为代表的模型Fine-tune完成文本匹配任务。 + +## 线上体验教程 + +- [使用seq2vec模块进行句子情感分类](https://aistudio.baidu.com/aistudio/projectdetail/1283423) + +- [如何将预训练模型Fine-tune下游任务](https://aistudio.baidu.com/aistudio/projectdetail/1294333) + +- [使用Bi-GRU+CRF完成快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) + +- [使用预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) + +- [使用Seq2Seq模型完成自动对联模型](https://aistudio.baidu.com/aistudio/projectdetail/1321118) + +- [使用预训练模型ERNIE-GEN实现智能写诗](https://aistudio.baidu.com/aistudio/projectdetail/1339888) + +- [使用TCN网络完成新冠疫情病例数预测](https://aistudio.baidu.com/aistudio/projectdetail/1290873) + +更多教程参见[PaddleNLP on AI Studio](https://aistudio.baidu.com/aistudio/personalcenter/thirdview/574995)。 diff --git a/PaddleNLP/examples/text_matching/sentence_transformers/README.md b/PaddleNLP/examples/text_matching/sentence_transformers/README.md index ad4547a9fa684e3563f188f2c096c2cd51dce107..f35931c7bf44884613c30a1b72c13862111f9617 100644 --- a/PaddleNLP/examples/text_matching/sentence_transformers/README.md +++ b/PaddleNLP/examples/text_matching/sentence_transformers/README.md @@ -39,7 +39,7 @@ PaddleNLP提供了丰富的预训练模型,并且可以便捷地获取PaddlePa 本项目针对中文文本匹配问题,开源了一系列模型,供用户可配置地使用: + BERT([Bidirectional Encoder Representations from Transformers](https://arxiv.org/abs/1810.04805))中文模型,简写`bert-base-chinese`, 其由12层Transformer网络组成。 -+ ERNIE([Enhanced Representation through Knowledge Integration](https://arxiv.org/pdf/1904.09223)),支持ERNIE 1.0中文模型(简写`ernie-1.0`)和ERNIE Tiny中文模型(简写`ernie-tiny`)。 ++ ERNIE([Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223)),支持ERNIE 1.0中文模型(简写`ernie-1.0`)和ERNIE Tiny中文模型(简写`ernie-tiny`)。 其中`ernie`由12层Transformer网络组成,`ernie-tiny`由3层Transformer网络组成。 + RoBERTa([A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692)),支持12层Transformer网络的`roberta-wwm-ext`。 @@ -195,3 +195,22 @@ Data: ['小蝌蚪找妈妈怎么样', '小蝌蚪找妈妈是谁画的'] Lab url = "https://arxiv.org/abs/2010.08240", } ``` + + +## 线上体验教程 + +- [使用seq2vec模块进行句子情感分类](https://aistudio.baidu.com/aistudio/projectdetail/1283423) + +- [如何将预训练模型Fine-tune下游任务](https://aistudio.baidu.com/aistudio/projectdetail/1294333) + +- [使用Bi-GRU+CRF完成快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) + +- [使用预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) + +- [使用Seq2Seq模型完成自动对联模型](https://aistudio.baidu.com/aistudio/projectdetail/1321118) + +- [使用预训练模型ERNIE-GEN实现智能写诗](https://aistudio.baidu.com/aistudio/projectdetail/1339888) + +- [使用TCN网络完成新冠疫情病例数预测](https://aistudio.baidu.com/aistudio/projectdetail/1290873) + +更多教程参见[PaddleNLP on AI Studio](https://aistudio.baidu.com/aistudio/personalcenter/thirdview/574995)。 diff --git a/PaddleNLP/examples/text_matching/simnet/README.md b/PaddleNLP/examples/text_matching/simnet/README.md index 26ccbbe0a1bf15dd0ad58d0e602d043a555423be..91c66b00626282a67e72d816f753d1e586ab5d73 100644 --- a/PaddleNLP/examples/text_matching/simnet/README.md +++ b/PaddleNLP/examples/text_matching/simnet/README.md @@ -164,3 +164,22 @@ Data: ['世界上什么东西最小', '世界上什么东西最小?'] Lab Data: ['光眼睛大就好看吗', '眼睛好看吗?'] Label: dissimilar Data: ['小蝌蚪找妈妈怎么样', '小蝌蚪找妈妈是谁画的'] Label: dissimilar ``` + + +## 线上体验教程 + +- [使用seq2vec模块进行句子情感分类](https://aistudio.baidu.com/aistudio/projectdetail/1283423) + +- [如何将预训练模型Fine-tune下游任务](https://aistudio.baidu.com/aistudio/projectdetail/1294333) + +- [使用Bi-GRU+CRF完成快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771) + +- [使用预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361) + +- [使用Seq2Seq模型完成自动对联模型](https://aistudio.baidu.com/aistudio/projectdetail/1321118) + +- [使用预训练模型ERNIE-GEN实现智能写诗](https://aistudio.baidu.com/aistudio/projectdetail/1339888) + +- [使用TCN网络完成新冠疫情病例数预测](https://aistudio.baidu.com/aistudio/projectdetail/1290873) + +更多教程参见[PaddleNLP on AI Studio](https://aistudio.baidu.com/aistudio/personalcenter/thirdview/574995)。 diff --git a/PaddleNLP/examples/time_series/README.md b/PaddleNLP/examples/time_series/README.md index 1f615fd9ac8b1ea35b0fa5f5147c0084ba3b0be2..651875911fc8e810ab3f20f60eb3581d640084c0 100644 --- a/PaddleNLP/examples/time_series/README.md +++ b/PaddleNLP/examples/time_series/README.md @@ -53,6 +53,8 @@ python predict.py --data_path time_series_covid19_confirmed_global.csv \ ``` -## 如何贡献代码 +## 线上教程体验 -如果你可以修复某个 issue 或者增加一个新功能,欢迎给我们提交 PR。如果对应的 PR 被接受了,我们将根据贡献的质量和难度 进行打分(0-5 分,越高越好)。如果你累计获得了 10 分,可以联系我们获得面试机会或为你写推荐信。 +我们为时间序列预测任务提供了线上教程,欢迎体验: + +* [使用TCN网络完成新冠疫情病例数预测](https://aistudio.baidu.com/aistudio/projectdetail/1290873) diff --git a/PaddleNLP/examples/time_series/covid-19_forecasting.ipynb b/PaddleNLP/examples/time_series/covid-19_forecasting.ipynb deleted file mode 100644 index 35a29a72155ccbd7c79b9c3526e4af05148286c0..0000000000000000000000000000000000000000 --- a/PaddleNLP/examples/time_series/covid-19_forecasting.ipynb +++ /dev/null @@ -1,589 +0,0 @@ -{ - "nbformat": 4, - "nbformat_minor": 2, - "metadata": { - "language_info": { - "name": "python", - "codemirror_mode": { - "name": "ipython", - "version": 3 - }, - "version": "3.6.10-final" - }, - "orig_nbformat": 2, - "file_extension": ".py", - "mimetype": "text/x-python", - "name": "python", - "npconvert_exporter": "python", - "pygments_lexer": "ipython3", - "version": 3, - "kernelspec": { - "name": "python3", - "display_name": "Python 3" - } - }, - "cells": [ - { - "source": [ - "# 使用PaddlePaddle完成新冠疫情病例数预测\n", - "\n", - "2019年12月以来,新冠疫情在全球肆虐,呈现大流行的特征。新型冠状病毒肺炎以发热、干咳、乏力等为主要表现,重症病例多在1周后出现呼吸困难,严重者快速进展为急性呼吸窘迫综合征、脓毒症休克、难以纠正的代谢性酸中毒和出凝血功能障碍及多器官功能衰竭等,对人们的健康造成了极其严重的威胁。同时,为抵御新冠病毒的扩散,不少国家和地区采取了封锁性防疫举措,全球经济复苏的进程因此受阻,政府债务不断上升。\n", - "\n", - "在这种背景下,各国人民都期盼着疫情的结束,早日恢复往常的生产、生活方式。本文关注到这一问题,结合约翰斯·霍普金斯大学发布的全球新冠肺炎实时统计数据,通过时间卷积神经网络对时间序列建模,实现预测未来病例数的目的。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "source": [ - "## 准备环境\n", - "\n", - "在开始建模之前,我们需要导入必要的包,同时为了更好地展示数据结果,我们在这里配置画图功能。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 1, - "metadata": {}, - "outputs": [ - { - "output_type": "stream", - "name": "stderr", - "text": "/mnt/qiujinxuan/PaddleNLP/paddlenlp/seq2vec/encoder.py:683: DeprecationWarning: invalid escape sequence \\s\n \"\"\"\n/mnt/qiujinxuan/PaddleNLP/paddlenlp/seq2vec/encoder.py:740: DeprecationWarning: invalid escape sequence \\s\n \"\"\"\n" - } - ], - "source": [ - "import os\n", - "import sys\n", - "\n", - "import paddle\n", - "import paddle.nn as nn\n", - "import numpy as np\n", - "import pandas as pd\n", - "import seaborn as sns\n", - "from pylab import rcParams\n", - "import matplotlib.pyplot as plt\n", - "from matplotlib import rc\n", - "from sklearn.preprocessing import MinMaxScaler\n", - "from pandas.plotting import register_matplotlib_converters\n", - "\n", - "sys.path.append(os.path.abspath(os.path.join(os.getcwd(), \"../..\")))\n", - "from paddlenlp.seq2vec import TCNEncoder\n", - "\n", - "\n", - "# config matplotlib\n", - "%matplotlib inline\n", - "%config InlineBackend.figure_format='retina'\n", - "sns.set(style='whitegrid', palette='muted', font_scale=1.2)\n", - "HAPPY_COLORS_PALETTE = [\"#01BEFE\", \"#FFDD00\", \"#FF7D00\", \"#FF006D\", \"#93D30C\", \"#8F00FF\"]\n", - "sns.set_palette(sns.color_palette(HAPPY_COLORS_PALETTE))\n", - "rcParams['figure.figsize'] = 14, 10\n", - "register_matplotlib_converters()" - ] - }, - { - "source": [ - "## 数据下载\n", - "\n", - "数据集由约翰·霍普金斯大学系统科学与工程中心提供,每日最新数据可以从https://github.com/CSSEGISandData/COVID-19 仓库中获取,我们在本例中提供了2020年11月24日下载的病例数据。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 2, - "metadata": {}, - "outputs": [], - "source": [ - "# !wget https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv" - ] - }, - { - "source": [ - "数据集中包含了国家、省份、纬度、经度以及从2020年1月22日至今的病例数等信息。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "source": [ - "## 数据预览\n", - "\n", - "数据集中包含了国家/地区、省份/州、纬度、经度、日期、病例数等信息。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 3, - "metadata": {}, - "outputs": [ - { - "output_type": "execute_result", - "data": { - "text/plain": " Province/State Country/Region Lat Long 1/22/20 1/23/20 \\\n0 NaN Afghanistan 33.93911 67.709953 0 0 \n1 NaN Albania 41.15330 20.168300 0 0 \n2 NaN Algeria 28.03390 1.659600 0 0 \n3 NaN Andorra 42.50630 1.521800 0 0 \n4 NaN Angola -11.20270 17.873900 0 0 \n\n 1/24/20 1/25/20 1/26/20 1/27/20 ... 11/13/20 11/14/20 11/15/20 \\\n0 0 0 0 0 ... 42969 43035 43240 \n1 0 0 0 0 ... 26701 27233 27830 \n2 0 0 0 0 ... 65975 66819 67679 \n3 0 0 0 0 ... 5725 5725 5872 \n4 0 0 0 0 ... 13228 13374 13451 \n\n 11/16/20 11/17/20 11/18/20 11/19/20 11/20/20 11/21/20 11/22/20 \n0 43403 43628 43851 44228 44443 44503 44706 \n1 28432 29126 29837 30623 31459 32196 32761 \n2 68589 69591 70629 71652 72755 73774 74862 \n3 5914 5951 6018 6066 6142 6207 6256 \n4 13615 13818 13922 14134 14267 14413 14493 \n\n[5 rows x 310 columns]", - "text/html": "
\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
Province/StateCountry/RegionLatLong1/22/201/23/201/24/201/25/201/26/201/27/20...11/13/2011/14/2011/15/2011/16/2011/17/2011/18/2011/19/2011/20/2011/21/2011/22/20
0NaNAfghanistan33.9391167.709953000000...42969430354324043403436284385144228444434450344706
1NaNAlbania41.1533020.168300000000...26701272332783028432291262983730623314593219632761
2NaNAlgeria28.033901.659600000000...65975668196767968589695917062971652727557377474862
3NaNAndorra42.506301.521800000000...5725572558725914595160186066614262076256
4NaNAngola-11.2027017.873900000000...13228133741345113615138181392214134142671441314493
\n

5 rows × 310 columns

\n
" - }, - "metadata": {}, - "execution_count": 3 - } - ], - "source": [ - "df_all = pd.read_csv('time_series_covid19_confirmed_global.csv')\n", - "df_all.head()" - ] - }, - { - "source": [ - "我们将对全世界的病例数进行预测,因此我们不需要关心具体国家的经纬度等信息,只需关注具体日期下的全球病例数即可。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 4, - "metadata": {}, - "outputs": [ - { - "output_type": "execute_result", - "data": { - "text/plain": "2020-01-22 555\n2020-01-23 654\n2020-01-24 941\n2020-01-25 1434\n2020-01-26 2118\ndtype: int64" - }, - "metadata": {}, - "execution_count": 4 - } - ], - "source": [ - "df = df_all.iloc[:, 4:]\n", - "daily_cases = df.sum(axis=0)\n", - "daily_cases.index = pd.to_datetime(daily_cases.index)\n", - "daily_cases.head()" - ] - }, - { - "cell_type": "code", - "execution_count": 5, - "metadata": {}, - "outputs": [ - { - "output_type": "display_data", - "data": { - "text/plain": "
", - "image/png": "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\n" - }, - "metadata": { - "image/png": { - "width": 826, - "height": 597 - } - } - } - ], - "source": [ - "plt.plot(daily_cases)\n", - "plt.title(\"Cumulative daily cases\");" - ] - }, - { - "source": [ - "为了提高样本时间序列的平稳性,我们继续取一阶差分。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 6, - "metadata": {}, - "outputs": [ - { - "output_type": "execute_result", - "data": { - "text/plain": "2020-01-22 555\n2020-01-23 99\n2020-01-24 287\n2020-01-25 493\n2020-01-26 684\ndtype: int64" - }, - "metadata": {}, - "execution_count": 6 - } - ], - "source": [ - "daily_cases = daily_cases.diff().fillna(daily_cases[0]).astype(np.int64)\n", - "daily_cases.head()" - ] - }, - { - "cell_type": "code", - "execution_count": 7, - "metadata": {}, - "outputs": [ - { - "output_type": "display_data", - "data": { - "text/plain": "
", - "image/png": "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\n" - }, - "metadata": { - "image/png": { - "width": 868, - "height": 597 - } - } - } - ], - "source": [ - "plt.plot(daily_cases)\n", - "plt.title(\"Daily cases\");" - ] - }, - { - "source": [ - "## 数据预处理\n", - "\n", - "首先划分数据集为训练集与验证集,我们取最后30天的数据作为测试集,其余作为训练集。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 8, - "metadata": {}, - "outputs": [ - { - "output_type": "stream", - "name": "stdout", - "text": "The number of the samples in train set is : 276\n" - } - ], - "source": [ - "TEST_DATA_SIZE = 30\n", - "\n", - "train_data = daily_cases[:-TEST_DATA_SIZE]\n", - "test_data = daily_cases[-TEST_DATA_SIZE:]\n", - "\n", - "print(\"The number of the samples in train set is : %i\"%train_data.shape[0])" - ] - }, - { - "source": [ - "为了提升模型收敛速度与性能,我们使用scikit-learn进行数据归一化。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 9, - "metadata": {}, - "outputs": [], - "source": [ - "scaler = MinMaxScaler()\n", - "train_data = scaler.fit_transform(np.expand_dims(train_data, axis=1)).astype('float32')\n", - "test_data = scaler.transform(np.expand_dims(test_data, axis=1)).astype('float32')" - ] - }, - { - "source": [ - "现在开始组建时间序列,可以用前10天的病例数预测当天的病例数。为了让测试集中的所有数据都能参与预测,我们将向测试集补充少量数据,这部分数据只会作为模型的输入。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 10, - "metadata": {}, - "outputs": [ - { - "output_type": "stream", - "name": "stdout", - "text": "The shape of x_train is: (267, 9, 1)\nThe shape of y_train is: (267, 1)\nThe shape of x_test is: (30, 9, 1)\nThe shape of y_test is: (30, 1)\n" - } - ], - "source": [ - "SEQ_LEN = 10\n", - "\n", - "def create_sequences(data, seq_length):\n", - " xs = []\n", - " ys = []\n", - "\n", - " for i in range(len(data)-seq_length+1):\n", - " x = data[i:i+seq_length-1]\n", - " y = data[i+seq_length-1]\n", - " xs.append(x)\n", - " ys.append(y)\n", - "\n", - " return np.array(xs), np.array(ys)\n", - "\n", - "x_train, y_train = create_sequences(train_data, SEQ_LEN)\n", - "test_data = np.concatenate((train_data[-SEQ_LEN+1:],test_data),axis=0)\n", - "x_test, y_test = create_sequences(test_data, SEQ_LEN)\n", - "\n", - "print(\"The shape of x_train is: %s\"%str(x_train.shape))\n", - "print(\"The shape of y_train is: %s\"%str(y_train.shape))\n", - "print(\"The shape of x_test is: %s\"%str(x_test.shape))\n", - "print(\"The shape of y_test is: %s\"%str(y_test.shape))" - ] - }, - { - "source": [ - "数据集处理完毕,将数据集封装到CovidDataset,以便模型训练、预测时调用。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 11, - "metadata": {}, - "outputs": [], - "source": [ - "class CovidDataset(paddle.io.Dataset):\n", - " def __init__(self, feature, label):\n", - " self.feature = feature\n", - " self.label = label\n", - " super(CovidDataset, self).__init__()\n", - "\n", - " def __len__(self):\n", - " return len(self.label)\n", - "\n", - " def __getitem__(self, index):\n", - " return [self.feature[index], self.label[index]]\n", - "\n", - "train_dataset = CovidDataset(x_train, y_train)\n", - "test_dataset = CovidDataset(x_test, y_test)\n" - ] - }, - { - "source": [ - "## 组网\n", - "\n", - "现在开始组建模型网络,我们采用时间卷积网络TCN作为特征提取器,将提取到的时序信息传送给全连接层获得最终的预测结果。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 12, - "metadata": {}, - "outputs": [], - "source": [ - "class TimeSeriesNetwork(nn.Layer):\n", - "\n", - " def __init__(self, input_size, next_k=1, num_channels=[64,128,256]):\n", - " super(TimeSeriesNetwork, self).__init__()\n", - "\n", - " self.last_num_channel = num_channels[-1]\n", - "\n", - " self.tcn = TCNEncoder(\n", - " input_size=input_size,\n", - " num_channels=num_channels,\n", - " kernel_size=2, \n", - " dropout=0.2\n", - " )\n", - "\n", - " self.linear = nn.Linear(in_features= self.last_num_channel, out_features=next_k)\n", - "\n", - " def forward(self, x):\n", - " tcn_out = self.tcn(x)\n", - " y_pred = self.linear(tcn_out)\n", - " return y_pred\n", - "\n", - "network = TimeSeriesNetwork(input_size=1)" - ] - }, - { - "source": [ - "## 定义优化器、损失函数\n", - "\n", - "在这里我们使用Adam优化器、均方差损失函数,为启动训练做最后的准备。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 13, - "metadata": {}, - "outputs": [], - "source": [ - "LR = 1e-3\n", - "\n", - "model = paddle.Model(network)\n", - "\n", - "optimizer = paddle.optimizer.Adam(\n", - " learning_rate=LR, parameters=model.parameters())\n", - "\n", - "loss = paddle.nn.MSELoss(reduction='sum')\n", - "\n", - "model.prepare(optimizer, loss)" - ] - }, - { - "source": [ - "## 训练\n", - "\n", - "配置必要的超参数,启动训练。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 14, - "metadata": {}, - "outputs": [ - { - "output_type": "stream", - "name": "stdout", - "text": "Epoch 1/100\nstep 8/8 [==============================] - loss: 0.3619 - 101ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/0\nEpoch 2/100\nstep 8/8 [==============================] - loss: 0.2746 - 53ms/step\nEpoch 3/100\nstep 8/8 [==============================] - loss: 0.0907 - 62ms/step\nEpoch 4/100\nstep 8/8 [==============================] - loss: 0.1740 - 55ms/step\nEpoch 5/100\nstep 8/8 [==============================] - loss: 0.1825 - 56ms/step\nEpoch 6/100\nstep 8/8 [==============================] - loss: 0.1334 - 56ms/step\nEpoch 7/100\nstep 8/8 [==============================] - loss: 0.1089 - 53ms/step\nEpoch 8/100\nstep 8/8 [==============================] - loss: 0.0948 - 57ms/step\nEpoch 9/100\nstep 8/8 [==============================] - loss: 0.1400 - 61ms/step\nEpoch 10/100\nstep 8/8 [==============================] - loss: 0.2118 - 52ms/step\nEpoch 11/100\nstep 8/8 [==============================] - loss: 0.1471 - 56ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/10\nEpoch 12/100\nstep 8/8 [==============================] - loss: 0.1984 - 54ms/step\nEpoch 13/100\nstep 8/8 [==============================] - loss: 0.1015 - 55ms/step\nEpoch 14/100\nstep 8/8 [==============================] - loss: 0.0796 - 58ms/step\nEpoch 15/100\nstep 8/8 [==============================] - loss: 0.0735 - 64ms/step\nEpoch 16/100\nstep 8/8 [==============================] - loss: 0.0608 - 62ms/step\nEpoch 17/100\nstep 8/8 [==============================] - loss: 0.1182 - 61ms/step\nEpoch 18/100\nstep 8/8 [==============================] - loss: 0.0943 - 58ms/step\nEpoch 19/100\nstep 8/8 [==============================] - loss: 0.0635 - 59ms/step\nEpoch 20/100\nstep 8/8 [==============================] - loss: 0.0990 - 53ms/step\nEpoch 21/100\nstep 8/8 [==============================] - loss: 0.0657 - 55ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/20\nEpoch 22/100\nstep 8/8 [==============================] - loss: 0.1203 - 56ms/step\nEpoch 23/100\nstep 8/8 [==============================] - loss: 0.0538 - 54ms/step\nEpoch 24/100\nstep 8/8 [==============================] - loss: 0.0576 - 57ms/step\nEpoch 25/100\nstep 8/8 [==============================] - loss: 0.1379 - 55ms/step\nEpoch 26/100\nstep 8/8 [==============================] - loss: 0.0858 - 54ms/step\nEpoch 27/100\nstep 8/8 [==============================] - loss: 0.0320 - 61ms/step\nEpoch 28/100\nstep 8/8 [==============================] - loss: 0.1156 - 58ms/step\nEpoch 29/100\nstep 8/8 [==============================] - loss: 0.1335 - 57ms/step\nEpoch 30/100\nstep 8/8 [==============================] - loss: 0.0953 - 58ms/step\nEpoch 31/100\nstep 8/8 [==============================] - loss: 0.0760 - 52ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/30\nEpoch 32/100\nstep 8/8 [==============================] - loss: 0.1796 - 54ms/step\nEpoch 33/100\nstep 8/8 [==============================] - loss: 0.1689 - 56ms/step\nEpoch 34/100\nstep 8/8 [==============================] - loss: 0.1669 - 52ms/step\nEpoch 35/100\nstep 8/8 [==============================] - loss: 0.0614 - 53ms/step\nEpoch 36/100\nstep 8/8 [==============================] - loss: 0.0823 - 56ms/step\nEpoch 37/100\nstep 8/8 [==============================] - loss: 0.0829 - 53ms/step\nEpoch 38/100\nstep 8/8 [==============================] - loss: 0.0622 - 55ms/step\nEpoch 39/100\nstep 8/8 [==============================] - loss: 0.0662 - 57ms/step\nEpoch 40/100\nstep 8/8 [==============================] - loss: 0.0642 - 52ms/step\nEpoch 41/100\nstep 8/8 [==============================] - loss: 0.0459 - 56ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/40\nEpoch 42/100\nstep 8/8 [==============================] - loss: 0.0460 - 56ms/step\nEpoch 43/100\nstep 8/8 [==============================] - loss: 0.1416 - 54ms/step\nEpoch 44/100\nstep 8/8 [==============================] - loss: 0.0700 - 57ms/step\nEpoch 45/100\nstep 8/8 [==============================] - loss: 0.0840 - 52ms/step\nEpoch 46/100\nstep 8/8 [==============================] - loss: 0.0784 - 54ms/step\nEpoch 47/100\nstep 8/8 [==============================] - loss: 0.1296 - 58ms/step\nEpoch 48/100\nstep 8/8 [==============================] - loss: 0.0487 - 53ms/step\nEpoch 49/100\nstep 8/8 [==============================] - loss: 0.0421 - 56ms/step\nEpoch 50/100\nstep 8/8 [==============================] - loss: 0.0534 - 57ms/step\nEpoch 51/100\nstep 8/8 [==============================] - loss: 0.0244 - 54ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/50\nEpoch 52/100\nstep 8/8 [==============================] - loss: 0.0725 - 57ms/step\nEpoch 53/100\nstep 8/8 [==============================] - loss: 0.1687 - 55ms/step\nEpoch 54/100\nstep 8/8 [==============================] - loss: 0.0693 - 55ms/step\nEpoch 55/100\nstep 8/8 [==============================] - loss: 0.0488 - 60ms/step\nEpoch 56/100\nstep 8/8 [==============================] - loss: 0.0737 - 54ms/step\nEpoch 57/100\nstep 8/8 [==============================] - loss: 0.0303 - 53ms/step\nEpoch 58/100\nstep 8/8 [==============================] - loss: 0.0457 - 60ms/step\nEpoch 59/100\nstep 8/8 [==============================] - loss: 0.0773 - 52ms/step\nEpoch 60/100\nstep 8/8 [==============================] - loss: 0.0385 - 54ms/step\nEpoch 61/100\nstep 8/8 [==============================] - loss: 0.1311 - 57ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/60\nEpoch 62/100\nstep 8/8 [==============================] - loss: 0.1149 - 53ms/step\nEpoch 63/100\nstep 8/8 [==============================] - loss: 0.1044 - 57ms/step\nEpoch 64/100\nstep 8/8 [==============================] - loss: 0.0670 - 55ms/step\nEpoch 65/100\nstep 8/8 [==============================] - loss: 0.0946 - 53ms/step\nEpoch 66/100\nstep 8/8 [==============================] - loss: 0.0565 - 57ms/step\nEpoch 67/100\nstep 8/8 [==============================] - loss: 0.1452 - 55ms/step\nEpoch 68/100\nstep 8/8 [==============================] - loss: 0.1813 - 53ms/step\nEpoch 69/100\nstep 8/8 [==============================] - loss: 0.0444 - 57ms/step\nEpoch 70/100\nstep 8/8 [==============================] - loss: 0.0626 - 55ms/step\nEpoch 71/100\nstep 8/8 [==============================] - loss: 0.0323 - 54ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/70\nEpoch 72/100\nstep 8/8 [==============================] - loss: 0.1469 - 59ms/step\nEpoch 73/100\nstep 8/8 [==============================] - loss: 0.0329 - 53ms/step\nEpoch 74/100\nstep 8/8 [==============================] - loss: 0.0430 - 54ms/step\nEpoch 75/100\nstep 8/8 [==============================] - loss: 0.0493 - 57ms/step\nEpoch 76/100\nstep 8/8 [==============================] - loss: 0.0693 - 52ms/step\nEpoch 77/100\nstep 8/8 [==============================] - loss: 0.0862 - 54ms/step\nEpoch 78/100\nstep 8/8 [==============================] - loss: 0.0316 - 57ms/step\nEpoch 79/100\nstep 8/8 [==============================] - loss: 0.0455 - 53ms/step\nEpoch 80/100\nstep 8/8 [==============================] - loss: 0.0341 - 54ms/step\nEpoch 81/100\nstep 8/8 [==============================] - loss: 0.1113 - 56ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/80\nEpoch 82/100\nstep 8/8 [==============================] - loss: 0.0367 - 52ms/step\nEpoch 83/100\nstep 8/8 [==============================] - loss: 0.0866 - 56ms/step\nEpoch 84/100\nstep 8/8 [==============================] - loss: 0.1026 - 54ms/step\nEpoch 85/100\nstep 8/8 [==============================] - loss: 0.0860 - 54ms/step\nEpoch 86/100\nstep 8/8 [==============================] - loss: 0.0386 - 57ms/step\nEpoch 87/100\nstep 8/8 [==============================] - loss: 0.0929 - 54ms/step\nEpoch 88/100\nstep 8/8 [==============================] - loss: 0.0560 - 54ms/step\nEpoch 89/100\nstep 8/8 [==============================] - loss: 0.0727 - 58ms/step\nEpoch 90/100\nstep 8/8 [==============================] - loss: 0.0682 - 54ms/step\nEpoch 91/100\nstep 8/8 [==============================] - loss: 0.0492 - 56ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/90\nEpoch 92/100\nstep 8/8 [==============================] - loss: 0.0604 - 56ms/step\nEpoch 93/100\nstep 8/8 [==============================] - loss: 0.0158 - 53ms/step\nEpoch 94/100\nstep 8/8 [==============================] - loss: 0.0412 - 57ms/step\nEpoch 95/100\nstep 8/8 [==============================] - loss: 0.0598 - 56ms/step\nEpoch 96/100\nstep 8/8 [==============================] - loss: 0.0185 - 54ms/step\nEpoch 97/100\nstep 8/8 [==============================] - loss: 0.0352 - 59ms/step\nEpoch 98/100\nstep 8/8 [==============================] - loss: 0.0192 - 58ms/step\nEpoch 99/100\nstep 8/8 [==============================] - loss: 0.0939 - 54ms/step\nEpoch 100/100\nstep 8/8 [==============================] - loss: 0.0525 - 61ms/step\nsave checkpoint at /mnt/qiujinxuan/PaddleNLP/examples/time_series/save_dir/final\n" - } - ], - "source": [ - "USE_GPU = False\n", - "TRAIN_EPOCH = 100\n", - "LOG_FREQ = 10\n", - "SAVE_DIR = os.path.join(os.getcwd(),\"save_dir\")\n", - "SAVE_FREQ = 10\n", - "\n", - "if USE_GPU:\n", - " paddle.set_device(\"gpu\")\n", - "else:\n", - " paddle.set_device(\"cpu\")\n", - "\n", - "model.fit(train_dataset, \n", - " batch_size=32,\n", - " drop_last=True,\n", - " epochs=TRAIN_EPOCH,\n", - " log_freq=LOG_FREQ,\n", - " save_dir=SAVE_DIR,\n", - " save_freq=SAVE_FREQ,\n", - " verbose=1\n", - " )" - ] - }, - { - "source": [ - "## 预测\n", - "\n", - "使用训练完毕的模型,对测试集中的日期对应的病例数进行预测。\n" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 15, - "metadata": {}, - "outputs": [ - { - "output_type": "stream", - "name": "stdout", - "text": "Predict begin...\nstep 30/30 [==============================] - 12ms/step \nPredict samples: 30\n" - } - ], - "source": [ - "preds = model.predict(\n", - " test_data=test_dataset\n", - " )" - ] - }, - { - "source": [ - "## 数据后处理\n", - "\n", - "将归一化的数据转换为原始数据,画出真实值对应的曲线和预测值对应的曲线。" - ], - "cell_type": "markdown", - "metadata": {} - }, - { - "cell_type": "code", - "execution_count": 16, - "metadata": {}, - "outputs": [], - "source": [ - "true_cases = scaler.inverse_transform(\n", - " np.expand_dims(y_test.flatten(), axis=0)\n", - ").flatten()\n", - "\n", - "predicted_cases = scaler.inverse_transform(\n", - " np.expand_dims(np.array(preds).flatten(), axis=0)\n", - ").flatten()" - ] - }, - { - "cell_type": "code", - "execution_count": 17, - "metadata": {}, - "outputs": [ - { - "output_type": "display_data", - "data": { - "text/plain": "
", - "image/png": "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\n" - }, - "metadata": { - "image/png": { - "width": 868, - "height": 580 - } - } - } - ], - "source": [ - "plt.plot(\n", - " daily_cases.index[:len(train_data)], \n", - " scaler.inverse_transform(train_data).flatten(),\n", - " label='Historical Daily Cases'\n", - ")\n", - "\n", - "plt.plot(\n", - " daily_cases.index[len(train_data):len(train_data) + len(true_cases)], \n", - " true_cases,\n", - " label='Real Daily Cases'\n", - ")\n", - "\n", - "plt.plot(\n", - " daily_cases.index[len(train_data):len(train_data) + len(true_cases)], \n", - " predicted_cases, \n", - " label='Predicted Daily Cases'\n", - ")\n", - "\n", - "plt.legend();" - ] - }, - { - "source": [ - "## 进一步优化模型\n", - "\n", - "从上图中,我们可以看到模型大体上预测到了病例数涨幅的升降情况,在具体数值上则出现了一些误差。读者可以发挥创造力,进一步提升模型的精度与功能,例如:\n", - "\n", - "1. 预测未来n天\n", - "\n", - "我们现在是用已知的9天病例数,预测第10天的病例数,我们可以将第10天的预测结果与前8天的真实病例数拼接,预测第11天的病例数,以此类推即可预测未来n天的病例数。\n", - "\n", - "2. 优化模型网络\n", - "\n", - "本文采用的是TCN模型,如果不考虑模型的速度性能,可以尝试LSTM, GRU, transformer等模型,进一步提升模型的拟合能力。\n", - "\n", - "3. 优化模型超参数\n", - "\n", - "本文没有对超参设置进行探索,读者可以探索设置更加合理的学习率,训练轮次,TCN通道数等。\n", - "\n", - "4. 考虑更多的数据特征\n", - "\n", - "本文只考虑了病例的日期,没有考虑政策、疫苗研制情况等具体环境的影响,读者可以搜集更多的新闻资料,加入更多的数据特征。\n", - "\n", - "我们欢迎您对本任务做出进一步的探索,如果您可以修复某个 issue 或者增加一个新功能,欢迎给我们提交 PR。如果对应的 PR 被接受了,我们将根据贡献的质量和难度 进行打分(0-5 分,越高越好)。如果你累计获得了 10 分,可以联系我们获得面试机会或为你写推荐信。" - ], - "cell_type": "markdown", - "metadata": {} - } - ] -} \ No newline at end of file diff --git a/PaddleNLP/paddlenlp/__init__.py b/PaddleNLP/paddlenlp/__init__.py index 98b2546fa63bb05fff519bb81a8bfb4a87b60b09..78baa6f82ae42b77d7bcef62f5dbaf9015d095b4 100644 --- a/PaddleNLP/paddlenlp/__init__.py +++ b/PaddleNLP/paddlenlp/__init__.py @@ -12,7 +12,7 @@ # See the License for the specific language governing permissions and # limitations under the License. -__version__ = '2.0.0a9' +__version__ = '2.0.0b0' from . import data from . import datasets