From 60dc4cc0146029b00dbc896b4b172a5cfdf1abab Mon Sep 17 00:00:00 2001 From: tink2123 Date: Mon, 17 Aug 2020 18:31:28 +0800 Subject: [PATCH] update doc --- README.md | 2 +- README_cn.md | 2 +- doc/doc_ch/recognition.md | 3 +++ 3 files changed, 5 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 32124165..d4c57d42 100644 --- a/README.md +++ b/README.md @@ -142,7 +142,7 @@ Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation r |RARE|MobileNetV3|83.32%|rec_mv3_tps_bilstm_attn|[Download link](https://paddleocr.bj.bcebos.com/rec_mv3_tps_bilstm_attn.tar)| |SRN|Resnet50_vd_fpn|88.33%|rec_r50fpn_vd_none_srn|[Download link](https://paddleocr.bj.bcebos.com/SRN/rec_r50fpn_vd_none_srn.tar)| -**Note:** SRN model uses data expansion method to expand the two training sets mentioned above, and the expanded data can be downloaded from [Baidu Drive](todo). +**Note:** SRN model uses data expansion method to expand the two training sets mentioned above, and the expanded data can be downloaded from [Baidu Drive](https://pan.baidu.com/s/1-HSZ-ZVdqBF2HaBZ5pRAKA), Extract the code:y3ry. The average accuracy of the two-stage training in the original paper is 89.74%, and that of one stage training in paddleocr is 88.33%. Both pre-trained weights can be downloaded [here](https://paddleocr.bj.bcebos.com/SRN/rec_r50fpn_vd_none_srn.tar). diff --git a/README_cn.md b/README_cn.md index c797097c..4ced5dfe 100644 --- a/README_cn.md +++ b/README_cn.md @@ -145,7 +145,7 @@ PaddleOCR开源的文本识别算法列表: |RARE|MobileNetV3|83.32%|rec_mv3_tps_bilstm_attn|[下载链接](https://paddleocr.bj.bcebos.com/rec_mv3_tps_bilstm_attn.tar)| |SRN|Resnet50_vd_fpn|88.33%|rec_r50fpn_vd_none_srn|[下载链接](https://paddleocr.bj.bcebos.com/SRN/rec_r50fpn_vd_none_srn.tar)| -**说明:** SRN模型使用了数据扰动方法对上述提到对两个训练集进行增广,增广后的数据可以在[百度网盘](todo)上下载。 +**说明:** SRN模型使用了数据扰动方法对上述提到对两个训练集进行增广,增广后的数据可以在[百度网盘](https://pan.baidu.com/s/1-HSZ-ZVdqBF2HaBZ5pRAKA)上下载,提取码: y3ry。 原始论文使用两阶段训练平均精度为89.74%,PaddleOCR中使用one-stage训练,平均精度为88.33%。两种预训练权重均在[下载链接](https://paddleocr.bj.bcebos.com/SRN/rec_r50fpn_vd_none_srn.tar)中。 使用[LSVT](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/doc/doc_ch/datasets.md#1icdar2019-lsvt)街景数据集根据真值将图crop出来30w数据,进行位置校准。此外基于LSVT语料生成500w合成数据训练中文模型,相关配置和预训练文件如下: diff --git a/doc/doc_ch/recognition.md b/doc/doc_ch/recognition.md index eda456c5..c554b9f1 100644 --- a/doc/doc_ch/recognition.md +++ b/doc/doc_ch/recognition.md @@ -18,6 +18,8 @@ ln -sf /train_data/dataset 若您本地没有数据集,可以在官网下载 [icdar2015](http://rrc.cvc.uab.es/?ch=4&com=downloads) 数据,用于快速验证。也可以参考[DTRB](https://github.com/clovaai/deep-text-recognition-benchmark#download-lmdb-dataset-for-traininig-and-evaluation-from-here),下载 benchmark 所需的lmdb格式数据集。 +如果希望复现SRN的论文指标,需要下载离线[增广数据](https://pan.baidu.com/s/1-HSZ-ZVdqBF2HaBZ5pRAKA),提取码: y3ry。增广数据是由MJSynth和SynthText做旋转和扰动得到的。数据下载完成后请解压到 {your_path}/PaddleOCR/train_data/data_lmdb_release/training/ 路径下。 + * 使用自己数据集: 若您希望使用自己的数据进行训练,请参考下文组织您的数据。 @@ -161,6 +163,7 @@ PaddleOCR支持训练和评估交替进行, 可以在 `configs/rec/rec_icdar15_t | rec_r34_vd_none_none_ctc.yml | Rosetta | Resnet34_vd | None | None | ctc | | rec_r34_vd_tps_bilstm_attn.yml | RARE | Resnet34_vd | tps | BiLSTM | attention | | rec_r34_vd_tps_bilstm_ctc.yml | STARNet | Resnet34_vd | tps | BiLSTM | ctc | +| rec_r50fpn_vd_none_srn.yml | SRN | Resnet50_fpn_vd | None | rnn | srn | 训练中文数据,推荐使用`rec_chinese_lite_train.yml`,如您希望尝试其他算法在中文数据集上的效果,请参考下列说明修改配置文件: -- GitLab