Skip to content

GitLab

"...composable_kernel_onnx.git" did not exist on "94bb1b48352bb3646a3e012513372c0692da9457"
Commit 2b3f89f0 authored by LDOUBLEV's avatar LDOUBLEV
Browse files

Merge branch 'dygraph' of https://github.com/PaddlePaddle/PaddleOCR into dygraph

parents 591b92e8 882e6e54
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 908 additions and 99 deletions
applications/多模态表单识别.md 0 → 100644
View file @ 2b3f89f0
This diff is collapsed.
configs/rec/ch_PP-OCRv3/ch_PP-OCRv3_rec.yml 0 → 100644
View file @ 2b3f89f0
Global:
debug: false
use_gpu: true
epoch_num: 500
log_smooth_window: 20
print_batch_step: 10
save_model_dir: ./output/rec_ppocr_v3
save_epoch_step: 3
eval_batch_step: [0, 2000]
cal_metric_during_train: true
pretrained_model:
checkpoints:
save_inference_dir:
use_visualdl: false
infer_img: doc/imgs_words/ch/word_1.jpg
character_dict_path: ppocr/utils/ppocr_keys_v1.txt
max_text_length: &max_text_length 25
infer_mode: false
use_space_char: true
distributed: true
save_res_path: ./output/rec/predicts_ppocrv3.txt
Optimizer:
name: Adam
beta1: 0.9
beta2: 0.999
lr:
name: Cosine
learning_rate: 0.001
warmup_epoch: 5
regularizer:
name: L2
factor: 3.0e-05
Architecture:
model_type: rec
algorithm: SVTR
Transform:
Backbone:
name: MobileNetV1Enhance
scale: 0.5
last_conv_stride: [1, 2]
last_pool_type: avg
Head:
name: MultiHead
head_list:
- CTCHead:
Neck:
name: svtr
dims: 64
depth: 2
hidden_dims: 120
use_guide: True
Head:
fc_decay: 0.00001
- SARHead:
enc_dim: 512
max_text_length: *max_text_length
Loss:
name: MultiLoss
loss_config_list:
- CTCLoss:
- SARLoss:
PostProcess:
name: CTCLabelDecode
Metric:
name: RecMetric
main_indicator: acc
ignore_space: True
Train:
dataset:
name: SimpleDataSet
data_dir: ./train_data/
ext_op_transform_idx: 1
label_file_list:
- ./train_data/train_list.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- RecConAug:
prob: 0.5
ext_data_num: 2
image_shape: [48, 320, 3]
- RecAug:
- MultiLabelEncode:
- RecResizeImg:
image_shape: [3, 48, 320]
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_sar
- length
- valid_ratio
loader:
shuffle: true
batch_size_per_card: 128
drop_last: true
num_workers: 4
Eval:
dataset:
name: SimpleDataSet
data_dir: ./train_data
label_file_list:
- ./train_data/val_list.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- MultiLabelEncode:
- RecResizeImg:
image_shape: [3, 48, 320]
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_sar
- length
- valid_ratio
loader:
shuffle: false
drop_last: false
batch_size_per_card: 128
num_workers: 4
configs/rec/ch_PP-OCRv3/ch_PP-OCRv3_rec_distillation.yml 0 → 100644
View file @ 2b3f89f0
Global:
debug: false
use_gpu: true
epoch_num: 800
log_smooth_window: 20
print_batch_step: 10
save_model_dir: ./output/rec_ppocr_v3_distillation
save_epoch_step: 3
eval_batch_step: [0, 2000]
cal_metric_during_train: true
pretrained_model:
checkpoints:
save_inference_dir:
use_visualdl: false
infer_img: doc/imgs_words/ch/word_1.jpg
character_dict_path: ppocr/utils/ppocr_keys_v1.txt
max_text_length: &max_text_length 25
infer_mode: false
use_space_char: true
distributed: true
save_res_path: ./output/rec/predicts_ppocrv3_distillation.txt
Optimizer:
name: Adam
beta1: 0.9
beta2: 0.999
lr:
name: Piecewise
decay_epochs : [700, 800]
values : [0.0005, 0.00005]
warmup_epoch: 5
regularizer:
name: L2
factor: 3.0e-05
Architecture:
model_type: &model_type "rec"
name: DistillationModel
algorithm: Distillation
Models:
Teacher:
pretrained:
freeze_params: false
return_all_feats: true
model_type: *model_type
algorithm: SVTR
Transform:
Backbone:
name: MobileNetV1Enhance
scale: 0.5
last_conv_stride: [1, 2]
last_pool_type: avg
Head:
name: MultiHead
head_list:
- CTCHead:
Neck:
name: svtr
dims: 64
depth: 2
hidden_dims: 120
use_guide: True
Head:
fc_decay: 0.00001
- SARHead:
enc_dim: 512
max_text_length: *max_text_length
Student:
pretrained:
freeze_params: false
return_all_feats: true
model_type: *model_type
algorithm: SVTR
Transform:
Backbone:
name: MobileNetV1Enhance
scale: 0.5
last_conv_stride: [1, 2]
last_pool_type: avg
Head:
name: MultiHead
head_list:
- CTCHead:
Neck:
name: svtr
dims: 64
depth: 2
hidden_dims: 120
use_guide: True
Head:
fc_decay: 0.00001
- SARHead:
enc_dim: 512
max_text_length: *max_text_length
Loss:
name: CombinedLoss
loss_config_list:
- DistillationDMLLoss:
weight: 1.0
act: "softmax"
use_log: true
model_name_pairs:
- ["Student", "Teacher"]
key: head_out
multi_head: True
dis_head: ctc
name: dml_ctc
- DistillationDMLLoss:
weight: 0.5
act: "softmax"
use_log: true
model_name_pairs:
- ["Student", "Teacher"]
key: head_out
multi_head: True
dis_head: sar
name: dml_sar
- DistillationDistanceLoss:
weight: 1.0
mode: "l2"
model_name_pairs:
- ["Student", "Teacher"]
key: backbone_out
- DistillationCTCLoss:
weight: 1.0
model_name_list: ["Student", "Teacher"]
key: head_out
multi_head: True
- DistillationSARLoss:
weight: 1.0
model_name_list: ["Student", "Teacher"]
key: head_out
multi_head: True
PostProcess:
name: DistillationCTCLabelDecode
model_name: ["Student", "Teacher"]
key: head_out
multi_head: True
Metric:
name: DistillationMetric
base_metric_name: RecMetric
main_indicator: acc
key: "Student"
ignore_space: True
Train:
dataset:
name: SimpleDataSet
data_dir: ./train_data/
ext_op_transform_idx: 1
label_file_list:
- ./train_data/train_list.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- RecConAug:
prob: 0.5
ext_data_num: 2
image_shape: [48, 320, 3]
- RecAug:
- MultiLabelEncode:
- RecResizeImg:
image_shape: [3, 48, 320]
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_sar
- length
- valid_ratio
loader:
shuffle: true
batch_size_per_card: 128
drop_last: true
num_workers: 4
Eval:
dataset:
name: SimpleDataSet
data_dir: ./train_data
label_file_list:
- ./train_data/val_list.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- MultiLabelEncode:
- RecResizeImg:
image_shape: [3, 48, 320]
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_sar
- length
- valid_ratio
loader:
shuffle: false
drop_last: false
batch_size_per_card: 128
num_workers: 4
configs/rec/rec_svtrnet.yml 0 → 100644
View file @ 2b3f89f0
Global:
use_gpu: True
epoch_num: 20
log_smooth_window: 20
print_batch_step: 10
save_model_dir: ./output/rec/svtr/
save_epoch_step: 1
# evaluation is run every 2000 iterations after the 0th iteration
eval_batch_step: [0, 2000]
cal_metric_during_train: True
pretrained_model:
checkpoints:
save_inference_dir:
use_visualdl: False
infer_img: doc/imgs_words_en/word_10.png
# for data or label process
character_dict_path:
character_type: en
max_text_length: 25
infer_mode: False
use_space_char: False
save_res_path: ./output/rec/predicts_svtr_tiny.txt
Optimizer:
name: AdamW
beta1: 0.9
beta2: 0.99
epsilon: 0.00000008
weight_decay: 0.05
no_weight_decay_name: norm pos_embed
one_dim_param_no_weight_decay: true
lr:
name: Cosine
learning_rate: 0.0005
warmup_epoch: 2
Architecture:
model_type: rec
algorithm: SVTR
Transform:
name: STN_ON
tps_inputsize: [32, 64]
tps_outputsize: [32, 100]
num_control_points: 20
tps_margins: [0.05,0.05]
stn_activation: none
Backbone:
name: SVTRNet
img_size: [32, 100]
out_char_num: 25
out_channels: 192
patch_merging: 'Conv'
embed_dim: [64, 128, 256]
depth: [3, 6, 3]
num_heads: [2, 4, 8]
mixer: ['Local','Local','Local','Local','Local','Local','Global','Global','Global','Global','Global','Global']
local_mixer: [[7, 11], [7, 11], [7, 11]]
last_stage: True
prenorm: false
Neck:
name: SequenceEncoder
encoder_type: reshape
Head:
name: CTCHead
Loss:
name: CTCLoss
PostProcess:
name: CTCLabelDecode
Metric:
name: RecMetric
main_indicator: acc
Train:
dataset:
name: LMDBDataSet
data_dir: ./train_data/data_lmdb_release/training/
transforms:
- DecodeImage: # load image
img_mode: BGR
channel_first: False
- CTCLabelEncode: # Class handling label
- RecResizeImg:
character_dict_path:
image_shape: [3, 64, 256]
padding: False
- KeepKeys:
keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
loader:
shuffle: True
batch_size_per_card: 512
drop_last: True
num_workers: 4
Eval:
dataset:
name: LMDBDataSet
data_dir: ./train_data/data_lmdb_release/validation/
transforms:
- DecodeImage: # load image
img_mode: BGR
channel_first: False
- CTCLabelEncode: # Class handling label
- RecResizeImg:
character_dict_path:
image_shape: [3, 64, 256]
padding: False
- KeepKeys:
keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
loader:
shuffle: False
drop_last: False
batch_size_per_card: 256
num_workers: 2
deploy/cpp_infer/readme.md
View file @ 2b3f89f0
- [Server-side C++ Inference](#server-side-c-inference) English | [简体中文](readme_ch.md)
- [1. Prepare the Environment](#1-prepare-the-environment)
- [Environment](#environment)
- [1.1 Compile OpenCV](#11-compile-opencv)
- [1.2 Compile or Download or the Paddle Inference Library](#12-compile-or-download-or-the-paddle-inference-library)
- [1.2.1 Direct download and installation](#121-direct-download-and-installation)
- [1.2.2 Compile the inference source code](#122-compile-the-inference-source-code)
- [2. Compile and Run the Demo](#2-compile-and-run-the-demo)
- [2.1 Export the inference model](#21-export-the-inference-model)
- [2.2 Compile PaddleOCR C++ inference demo](#22-compile-paddleocr-c-inference-demo)
- [Run the demo](#run-the-demo)
- [1. det+cls+rec：](#1-detclsrec)
- [2. det+rec：](#2-detrec)
- [3. det](#3-det)
- [4. cls+rec：](#4-clsrec)
- [5. rec](#5-rec)
- [6. cls](#6-cls)
- [3. FAQ](#3-faq)
# Server-side C++ Inference # Server-side C++ Inference
This chapter introduces the C++ deployment steps of the PaddleOCR model. The corresponding Python predictive deployment method refers to [document](../../doc/doc_ch/inference.md). - [1. Prepare the Environment](#1)
C++ is better than python in terms of performance. Therefore, in CPU and GPU deployment scenarios, C++ deployment is mostly used. - [1.1 Environment](#11)
- [1.2 Compile OpenCV](#12)
- [1.3 Compile or Download or the Paddle Inference Library](#13)
- [2. Compile and Run the Demo](#2)
- [2.1 Export the inference model](#21)
- [2.2 Compile PaddleOCR C++ inference demo](#22)
- [2.3 Run the demo](#23)
- [3. FAQ](#3)
This chapter introduces the C++ deployment steps of the PaddleOCR model. C++ is better than Python in terms of performance. Therefore, in CPU and GPU deployment scenarios, C++ deployment is mostly used.
This section will introduce how to configure the C++ environment and deploy PaddleOCR in Linux (CPU\GPU) environment. For Windows deployment please refer to [Windows](./docs/windows_vs2019_build.md) compilation guidelines. This section will introduce how to configure the C++ environment and deploy PaddleOCR in Linux (CPU\GPU) environment. For Windows deployment please refer to [Windows](./docs/windows_vs2019_build.md) compilation guidelines.
<a name="1"></a>
## 1. Prepare the Environment ## 1. Prepare the Environment
### Environment <a name="11"></a>
### 1.1 Environment
- Linux, docker is recommended. - Linux, docker is recommended.
- Windows. - Windows.
### 1.1 Compile OpenCV <a name="12"></a>
### 1.2 Compile OpenCV
* First of all, you need to download the source code compiled package in the Linux environment from the OpenCV official website. Taking OpenCV 3.4.7 as an example, the download command is as follows. * First of all, you need to download the source code compiled package in the Linux environment from the OpenCV official website. Taking OpenCV 3.4.7 as an example, the download command is as follows.
...@@ -92,11 +88,12 @@ opencv3/ ...@@ -92,11 +88,12 @@ opencv3/
|-- share |-- share
``` ```
### 1.2 Compile or Download or the Paddle Inference Library <a name="13"></a>
### 1.3 Compile or Download or the Paddle Inference Library
* There are 2 ways to obtain the Paddle inference library, described in detail below. * There are 2 ways to obtain the Paddle inference library, described in detail below.
#### 1.2.1 Direct download and installation #### 1.3.1 Direct download and installation
[Paddle inference library official website](https://paddleinference.paddlepaddle.org.cn/user_guides/download_lib.html#linux). You can review and select the appropriate version of the inference library on the official website. [Paddle inference library official website](https://paddleinference.paddlepaddle.org.cn/user_guides/download_lib.html#linux). You can review and select the appropriate version of the inference library on the official website.
...@@ -109,7 +106,7 @@ tar -xf paddle_inference.tgz ...@@ -109,7 +106,7 @@ tar -xf paddle_inference.tgz
Finally you will see the the folder of `paddle_inference/` in the current path. Finally you will see the the folder of `paddle_inference/` in the current path.
#### 1.2.2 Compile the inference source code #### 1.3.2 Compile the inference source code
* If you want to get the latest Paddle inference library features, you can download the latest code from Paddle GitHub repository and compile the inference library from the source code. It is recommended to download the inference library with paddle version greater than or equal to 2.0.1. * If you want to get the latest Paddle inference library features, you can download the latest code from Paddle GitHub repository and compile the inference library from the source code. It is recommended to download the inference library with paddle version greater than or equal to 2.0.1.
* You can refer to [Paddle inference library] (https://www.paddlepaddle.org.cn/documentation/docs/en/advanced_guide/inference_deployment/inference/build_and_install_lib_en.html) to get the Paddle source code from GitHub, and then compile To generate the latest inference library. The method of using git to access the code is as follows. * You can refer to [Paddle inference library] (https://www.paddlepaddle.org.cn/documentation/docs/en/advanced_guide/inference_deployment/inference/build_and_install_lib_en.html) to get the Paddle source code from GitHub, and then compile To generate the latest inference library. The method of using git to access the code is as follows.
...@@ -155,8 +152,10 @@ build/paddle_inference_install_dir/ ...@@ -155,8 +152,10 @@ build/paddle_inference_install_dir/
`paddle` is the Paddle library required for C++ prediction later, and `version.txt` contains the version information of the current inference library. `paddle` is the Paddle library required for C++ prediction later, and `version.txt` contains the version information of the current inference library.
<a name="2"></a>
## 2. Compile and Run the Demo ## 2. Compile and Run the Demo
<a name="21"></a>
### 2.1 Export the inference model ### 2.1 Export the inference model
* You can refer to [Model inference](../../doc/doc_ch/inference.md) and export the inference model. After the model is exported, assuming it is placed in the `inference` directory, the directory structure is as follows. * You can refer to [Model inference](../../doc/doc_ch/inference.md) and export the inference model. After the model is exported, assuming it is placed in the `inference` directory, the directory structure is as follows.
...@@ -175,9 +174,9 @@ inference/ ...@@ -175,9 +174,9 @@ inference/
``` ```
<a name="22"></a>
### 2.2 Compile PaddleOCR C++ inference demo ### 2.2 Compile PaddleOCR C++ inference demo
* The compilation commands are as follows. The addresses of Paddle C++ inference library, opencv and other Dependencies need to be replaced with the actual addresses on your own machines. * The compilation commands are as follows. The addresses of Paddle C++ inference library, opencv and other Dependencies need to be replaced with the actual addresses on your own machines.
```shell ```shell
...@@ -201,7 +200,9 @@ or the generated Paddle inference library path (`build/paddle_inference_install_ ...@@ -201,7 +200,9 @@ or the generated Paddle inference library path (`build/paddle_inference_install_
* After the compilation is completed, an executable file named `ppocr` will be generated in the `build` folder. * After the compilation is completed, an executable file named `ppocr` will be generated in the `build` folder.
### Run the demo <a name="23"></a>
### 2.3 Run the demo
Execute the built executable file: Execute the built executable file:
```shell ```shell
./build/ppocr [--param1] [--param2] [...] ./build/ppocr [--param1] [--param2] [...]
...@@ -342,6 +343,7 @@ The detection visualized image saved in ./output//12.jpg ...@@ -342,6 +343,7 @@ The detection visualized image saved in ./output//12.jpg
``` ```
<a name="3"></a>
## 3. FAQ ## 3. FAQ
1. Encountered the error `unable to access 'https://github.com/LDOUBLEV/AutoLog.git/': gnutls_handshake() failed: The TLS connection was non-properly terminated.`, change the github address in `deploy/cpp_infer/external-cmake/auto-log.cmake` to the https://gitee.com/Double_V/AutoLog address. 1. Encountered the error `unable to access 'https://github.com/LDOUBLEV/AutoLog.git/': gnutls_handshake() failed: The TLS connection was non-properly terminated.`, change the github address in `deploy/cpp_infer/external-cmake/auto-log.cmake` to the https://gitee.com/Double_V/AutoLog address.
deploy/cpp_infer/readme_ch.md
View file @ 2b3f89f0
- [服务器端C++预测](#服务器端c预测) [English](readme.md) | 简体中文
- [1. 准备环境](#1-准备环境)
- [1.0 运行准备](#10-运行准备)
- [1.1 编译opencv库](#11-编译opencv库)
- [1.2 下载或者编译Paddle预测库](#12-下载或者编译paddle预测库)
- [1.2.1 直接下载安装](#121-直接下载安装)
- [1.2.2 预测库源码编译](#122-预测库源码编译)
- [2 开始运行](#2-开始运行)
- [2.1 将模型导出为inference model](#21-将模型导出为inference-model)
- [2.2 编译PaddleOCR C++预测demo](#22-编译paddleocr-c预测demo)
- [2.3 运行demo](#23-运行demo)
- [1. 检测+分类+识别：](#1-检测分类识别)
- [2. 检测+识别：](#2-检测识别)
- [3. 检测：](#3-检测)
- [4. 分类+识别：](#4-分类识别)
- [5. 识别：](#5-识别)
- [6. 分类：](#6-分类)
- [3. FAQ](#3-faq)
# 服务器端C++预测 # 服务器端C++预测
本章节介绍PaddleOCR 模型的的C++部署方法，与之对应的python预测部署方式参考[文档](../../doc/doc_ch/inference.md)。 - [1. 准备环境](#1)
C++在性能计算上优于python，因此，在大多数CPU、GPU部署场景，多采用C++的部署方式，本节将介绍如何在Linux\Windows (CPU\GPU)环境下配置C++环境并完成 - [1.1 运行准备](#11)
PaddleOCR模型部署。 - [1.2 编译opencv库](#12)
- [1.3 下载或者编译Paddle预测库](#13)
- [2 开始运行](#2)
- [2.1 准备模型](#21)
- [2.2 编译PaddleOCR C++预测demo](#22)
- [2.3 运行demo](#23)
- [3. FAQ](#3)
本章节介绍PaddleOCR 模型的的C++部署方法。C++在性能计算上优于Python，因此，在大多数CPU、GPU部署场景，多采用C++的部署方式，本节将介绍如何在Linux\Windows (CPU\GPU)环境下配置C++环境并完成PaddleOCR模型部署。
<a name="1"></a> <a name="1"></a>
## 1. 准备环境 ## 1. 准备环境
<a name="10"></a> <a name="11"></a>
### 1.0 运行准备 ### 1.1 运行准备
- Linux环境，推荐使用docker。 - Linux环境，推荐使用docker。
- Windows环境。 - Windows环境。
* 该文档主要介绍基于Linux环境的PaddleOCR C++预测流程，如果需要在Windows下基于预测库进行C++预测，具体编译方法请参考[Windows下编译教程](./docs/windows_vs2019_build.md) * 该文档主要介绍基于Linux环境的PaddleOCR C++预测流程，如果需要在Windows下基于预测库进行C++预测，具体编译方法请参考[Windows下编译教程](./docs/windows_vs2019_build.md)
<a name="11"></a> <a name="12"></a>
### 1.1 编译opencv库 ### 1.2 编译opencv库
* 首先需要从opencv官网上下载在Linux环境下源码编译的包，以opencv3.4.7为例，下载命令如下。 * 首先需要从opencv官网上下载在Linux环境下源码编译的包，以opencv3.4.7为例，下载命令如下。
...@@ -103,35 +94,38 @@ opencv3/ ...@@ -103,35 +94,38 @@ opencv3/
|-- share |-- share
``` ```
<a name="12"></a> <a name="13"></a>
### 1.2 下载或者编译Paddle预测库
* 有2种方式获取Paddle预测库，下面进行详细介绍。 ### 1.3 下载或者编译Paddle预测库
可以选择直接下载安装或者从源码编译，下文分别进行具体说明。
#### 1.2.1 直接下载安装 <a name="131"></a>
#### 1.3.1 直接下载安装
* [Paddle预测库官网](https://paddleinference.paddlepaddle.org.cn/user_guides/download_lib.html#linux) 上提供了不同cuda版本的Linux预测库，可以在官网查看并选择合适的预测库版本（*建议选择paddle版本>=2.0.1版本的预测库* ）。 [Paddle预测库官网](https://paddleinference.paddlepaddle.org.cn/user_guides/download_lib.html#linux) 上提供了不同cuda版本的Linux预测库，可以在官网查看并选择合适的预测库版本（*建议选择paddle版本>=2.0.1版本的预测库* ）。
* 下载之后使用下面的方法解压。 下载之后解压:
``` ```shell
tar -xf paddle_inference.tgz tar -xf paddle_inference.tgz
``` ```
最终会在当前的文件夹中生成`paddle_inference/`的子文件夹。 最终会在当前的文件夹中生成`paddle_inference/`的子文件夹。
#### 1.2.2 预测库源码编译 <a name="132"></a>
* 如果希望获取最新预测库特性，可以从Paddle github上克隆最新代码，源码编译预测库。 #### 1.3.2 预测库源码编译
* 可以参考[Paddle预测库安装编译说明](https://www.paddlepaddle.org.cn/documentation/docs/zh/2.0/guides/05_inference_deployment/inference/build_and_install_lib_cn.html#congyuanmabianyi) 的说明，从github上获取Paddle代码，然后进行编译，生成最新的预测库。使用git获取代码方法如下。
如果希望获取最新预测库特性，可以从github上克隆最新Paddle代码进行编译，生成最新的预测库。
* 使用git获取代码:
```shell ```shell
git clone https://github.com/PaddlePaddle/Paddle.git git clone https://github.com/PaddlePaddle/Paddle.git
git checkout develop git checkout develop
``` ```
* 进入Paddle目录后，编译方法如下。 * 进入Paddle目录，进行编译:
```shell ```shell
rm -rf build rm -rf build
...@@ -151,7 +145,7 @@ make -j ...@@ -151,7 +145,7 @@ make -j
make inference_lib_dist make inference_lib_dist
``` ```
更多编译参数选项介绍可以参考[文档说明](https://www.paddlepaddle.org.cn/documentation/docs/zh/2.0/guides/05_inference_deployment/inference/build_and_install_lib_cn.html#congyuanmabianyi)。 更多编译参数选项介绍可以参考[Paddle预测库编译文档](https://www.paddlepaddle.org.cn/documentation/docs/zh/2.0/guides/05_inference_deployment/inference/build_and_install_lib_cn.html#congyuanmabianyi)。
* 编译完成之后，可以在`build/paddle_inference_install_dir/`文件下看到生成了以下文件及文件夹。 * 编译完成之后，可以在`build/paddle_inference_install_dir/`文件下看到生成了以下文件及文件夹。
...@@ -168,13 +162,13 @@ build/paddle_inference_install_dir/ ...@@ -168,13 +162,13 @@ build/paddle_inference_install_dir/
<a name="2"></a> <a name="2"></a>
## 2 开始运行 ## 2. 开始运行
<a name="21"></a> <a name="21"></a>
### 2.1 将模型导出为inference model ### 2.1 准备模型
* 可以参考[模型预测章节](../../doc/doc_ch/inference.md)，导出inference model，用于模型预测。模型导出之后，假设放在`inference`目录下，则目录结构如下。 直接下载PaddleOCR提供的推理模型，或者参考[模型预测章节](../../doc/doc_ch/inference_ppocr.md)，将训练好的模型导出为推理模型。模型导出之后，假设放在`inference`目录下，则目录结构如下。
``` ```
inference/ inference/
...@@ -193,13 +187,13 @@ inference/ ...@@ -193,13 +187,13 @@ inference/
### 2.2 编译PaddleOCR C++预测demo ### 2.2 编译PaddleOCR C++预测demo
* 编译命令如下，其中Paddle C++预测库、opencv等其他依赖库的地址需要换成自己机器上的实际地址。 编译命令如下，其中Paddle C++预测库、opencv等其他依赖库的地址需要换成自己机器上的实际地址。
```shell ```shell
sh tools/build.sh sh tools/build.sh
``` ```
* 具体的，需要修改`tools/build.sh`中环境路径，相关内容如下： 具体的，需要修改`tools/build.sh`中环境路径，相关内容如下：
```shell ```shell
OPENCV_DIR=your_opencv_dir OPENCV_DIR=your_opencv_dir
...@@ -211,12 +205,14 @@ CUDNN_LIB_DIR=/your_cudnn_lib_dir ...@@ -211,12 +205,14 @@ CUDNN_LIB_DIR=/your_cudnn_lib_dir
其中，`OPENCV_DIR`为opencv编译安装的地址；`LIB_DIR`为下载(`paddle_inference`文件夹)或者编译生成的Paddle预测库地址(`build/paddle_inference_install_dir`文件夹)；`CUDA_LIB_DIR`为cuda库文件地址，在docker中为`/usr/local/cuda/lib64`；`CUDNN_LIB_DIR`为cudnn库文件地址，在docker中为`/usr/lib/x86_64-linux-gnu/`。**注意：以上路径都写绝对路径，不要写相对路径。** 其中，`OPENCV_DIR`为opencv编译安装的地址；`LIB_DIR`为下载(`paddle_inference`文件夹)或者编译生成的Paddle预测库地址(`build/paddle_inference_install_dir`文件夹)；`CUDA_LIB_DIR`为cuda库文件地址，在docker中为`/usr/local/cuda/lib64`；`CUDNN_LIB_DIR`为cudnn库文件地址，在docker中为`/usr/lib/x86_64-linux-gnu/`。**注意：以上路径都写绝对路径，不要写相对路径。**
* 编译完成之后，会在`build`文件夹下生成一个名为`ppocr`的可执行文件。 编译完成之后，会在`build`文件夹下生成一个名为`ppocr`的可执行文件。
<a name="23"></a> <a name="23"></a>
### 2.3 运行demo ### 2.3 运行demo
本demo支持系统串联调用，也支持单个功能的调用，如，只使用检测或识别功能。
运行方式： 运行方式：
```shell ```shell
./build/ppocr [--param1] [--param2] [...] ./build/ppocr [--param1] [--param2] [...]
...@@ -354,6 +350,7 @@ predict img: ../../doc/imgs/12.jpg ...@@ -354,6 +350,7 @@ predict img: ../../doc/imgs/12.jpg
The detection visualized image saved in ./output//12.jpg The detection visualized image saved in ./output//12.jpg
``` ```
<a name="3"></a>
## 3. FAQ ## 3. FAQ
1. 遇到报错 `unable to access 'https://github.com/LDOUBLEV/AutoLog.git/': gnutls_handshake() failed: The TLS connection was non-properly terminated.`， 将 `deploy/cpp_infer/external-cmake/auto-log.cmake` 中的github地址改为 https://gitee.com/Double_V/AutoLog 地址即可。 1. 遇到报错 `unable to access 'https://github.com/LDOUBLEV/AutoLog.git/': gnutls_handshake() failed: The TLS connection was non-properly terminated.`， 将 `deploy/cpp_infer/external-cmake/auto-log.cmake` 中的github地址改为 https://gitee.com/Double_V/AutoLog 地址即可。
deploy/slim/quantization/quant.py
View file @ 2b3f89f0
...@@ -137,7 +137,7 @@ def main(config, device, logger, vdl_writer): ...@@ -137,7 +137,7 @@ def main(config, device, logger, vdl_writer):
config['Optimizer'], config['Optimizer'],
epochs=config['Global']['epoch_num'], epochs=config['Global']['epoch_num'],
step_each_epoch=len(train_dataloader), step_each_epoch=len(train_dataloader),
parameters=model.parameters()) model=model)
# resume PACT training process # resume PACT training process
if config["Global"]["checkpoints"] is not None: if config["Global"]["checkpoints"] is not None:
......
doc/doc_ch/algorithm_det_db.md
View file @ 2b3f89f0
...@@ -47,13 +47,13 @@ ...@@ -47,13 +47,13 @@
### 4.1 Python推理 ### 4.1 Python推理
首先将DB文本检测训练过程中保存的模型，转换成inference model。以基于Resnet50_vd骨干网络，在ICDAR2015英文数据集训练的模型为例（ [模型下载地址](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_db_v2.0_train.tar) )，可以使用如下命令进行转换： 首先将DB文本检测训练过程中保存的模型，转换成inference model。以基于Resnet50_vd骨干网络，在ICDAR2015英文数据集训练的模型为例（ [模型下载地址](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_db_v2.0_train.tar) )，可以使用如下命令进行转换：
``` ```shell
python3 tools/export_model.py -c configs/det/det_r50_vd_db.yml -o Global.pretrained_model=./det_r50_vd_db_v2.0_train/best_accuracy Global.save_inference_dir=./inference/det_db python3 tools/export_model.py -c configs/det/det_r50_vd_db.yml -o Global.pretrained_model=./det_r50_vd_db_v2.0_train/best_accuracy Global.save_inference_dir=./inference/det_db
``` ```
DB文本检测模型推理，可以执行如下命令： DB文本检测模型推理，可以执行如下命令：
``` ```shell
python3 tools/infer/predict_det.py --image_dir="./doc/imgs_en/img_10.jpg" --det_model_dir="./inference/det_db/" python3 tools/infer/predict_det.py --image_dir="./doc/imgs_en/img_10.jpg" --det_model_dir="./inference/det_db/"
``` ```
...@@ -65,15 +65,20 @@ python3 tools/infer/predict_det.py --image_dir="./doc/imgs_en/img_10.jpg" --det_ ...@@ -65,15 +65,20 @@ python3 tools/infer/predict_det.py --image_dir="./doc/imgs_en/img_10.jpg" --det_
<a name="4-2"></a> <a name="4-2"></a>
### 4.2 C++推理 ### 4.2 C++推理
敬请期待
准备好推理模型后，参考[cpp infer](../../deploy/cpp_infer/)教程进行操作即可。
<a name="4-3"></a> <a name="4-3"></a>
### 4.3 Serving服务化部署 ### 4.3 Serving服务化部署
敬请期待
准备好推理模型后，参考[pdserving](../../deploy/pdserving/)教程进行Serving服务化部署，包括Python Serving和C++ Serving两种模式。
<a name="4-4"></a> <a name="4-4"></a>
### 4.4 更多推理部署 ### 4.4 更多推理部署
敬请期待
DB模型还支持以下推理部署方式：
- Paddle2ONNX推理：准备好推理模型后，参考[paddle2onnx](../../deploy/paddle2onnx/)教程操作。
<a name="5"></a> <a name="5"></a>
## 5. FAQ ## 5. FAQ
......
doc/doc_en/algorithm_det_db_en.md
View file @ 2b3f89f0
...@@ -14,4 +14,86 @@ ...@@ -14,4 +14,86 @@
- [5. FAQ](#5) - [5. FAQ](#5)
<a name="1"></a> <a name="1"></a>
## 1. Introduction ## 1. Introduction
\ No newline at end of file
Paper:
> [Real-time Scene Text Detection with Differentiable Binarization](https://arxiv.org/abs/1911.08947)
> Liao, Minghui and Wan, Zhaoyi and Yao, Cong and Chen, Kai and Bai, Xiang
> AAAI, 2020
On the ICDAR2015 dataset, the text detection result is as follows:
|Model|Backbone|Configuration|Precision|Recall|Hmean|Download|
| --- | --- | --- | --- | --- | --- | --- |
|DB|ResNet50_vd|configs/det/det_r50_vd_db.yml|86.41%|78.72%|82.38%|[trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_db_v2.0_train.tar)|
|DB|MobileNetV3|configs/det/det_mv3_db.yml|77.29%|73.08%|75.12%|[trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_mv3_db_v2.0_train.tar)|
<a name="2"></a>
## 2. Environment
Please prepare your environment referring to [prepare the environment](./environment_en.md) and [clone the repo](./clone_en.md).
<a name="3"></a>
## 3. Model Training / Evaluation / Prediction
Please refer to [text detection training tutorial](./detection_en.md). PaddleOCR has modularized the code structure, so that you only need to **replace the configuration file** to train different detection models.
<a name="4"></a>
## 4. Inference and Deployment
<a name="4-1"></a>
### 4.1 Python Inference
First, convert the model saved in the DB text detection training process into an inference model. Taking the model based on the Resnet50_vd backbone network and trained on the ICDAR2015 English dataset as example ([model download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_db_v2.0_train.tar)), you can use the following command to convert:
```shell
python3 tools/export_model.py -c configs/det/det_r50_vd_db.yml -o Global.pretrained_model=./det_r50_vd_db_v2.0_train/best_accuracy Global.save_inference_dir=./inference/det_db
```
DB text detection model inference, you can execute the following command:
```shell
python3 tools/infer/predict_det.py --image_dir="./doc/imgs_en/img_10.jpg" --det_model_dir="./inference/det_db/"
```
The visualized text detection results are saved to the `./inference_results` folder by default, and the name of the result file is prefixed with 'det_res'. Examples of results are as follows:
![](../imgs_results/det_res_img_10_db.jpg)
**Note**: Since the ICDAR2015 dataset has only 1,000 training images, mainly for English scenes, the above model has very poor detection result on Chinese text images.
<a name="4-2"></a>
### 4.2 C++ Inference
With the inference model prepared, refer to the [cpp infer](../../deploy/cpp_infer/) tutorial for C++ inference.
<a name="4-3"></a>
### 4.3 Serving
With the inference model prepared, refer to the [pdserving](../../deploy/pdserving/) tutorial for service deployment by Paddle Serving.
<a name="4-4"></a>
### 4.4 More
More deployment schemes supported for DB:
- Paddle2ONNX: with the inference model prepared, please refer to the [paddle2onnx](../../deploy/paddle2onnx/) tutorial.
<a name="5"></a>
## 5. FAQ
## Citation
```bibtex
@inproceedings{liao2020real,
title={Real-time scene text detection with differentiable binarization},
author={Liao, Minghui and Wan, Zhaoyi and Yao, Cong and Chen, Kai and Bai, Xiang},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
volume={34},
number={07},
pages={11474--11481},
year={2020}
}
```
\ No newline at end of file
doc/joinus.PNG
View replaced file @ 591b92e8 View file @ 2b3f89f0
doc/joinus.PNG

201 KB | W: | H:

doc/joinus.PNG

200 KB | W: | H:

doc/joinus.PNG
doc/joinus.PNG
doc/joinus.PNG
doc/joinus.PNG
  • 2-up
  • Swipe
  • Onion skin
ppocr/data/imaug/__init__.py
View file @ 2b3f89f0
...@@ -22,8 +22,8 @@ from .make_shrink_map import MakeShrinkMap ...@@ -22,8 +22,8 @@ from .make_shrink_map import MakeShrinkMap
from .random_crop_data import EastRandomCropData, RandomCropImgMask from .random_crop_data import EastRandomCropData, RandomCropImgMask
from .make_pse_gt import MakePseGt from .make_pse_gt import MakePseGt
from .rec_img_aug import RecAug, RecResizeImg, ClsResizeImg, \ from .rec_img_aug import RecAug, RecConAug, RecResizeImg, ClsResizeImg, \
SRNRecResizeImg, NRTRRecResizeImg, SARRecResizeImg, PRENResizeImg SRNRecResizeImg, NRTRRecResizeImg, SARRecResizeImg, PRENResizeImg, SVTRRecResizeImg
from .randaugment import RandAugment from .randaugment import RandAugment
from .copy_paste import CopyPaste from .copy_paste import CopyPaste
from .ColorJitter import ColorJitter from .ColorJitter import ColorJitter
......
ppocr/data/imaug/label_ops.py
View file @ 2b3f89f0
...@@ -22,6 +22,7 @@ import numpy as np ...@@ -22,6 +22,7 @@ import numpy as np
import string import string
from shapely.geometry import LineString, Point, Polygon from shapely.geometry import LineString, Point, Polygon
import json import json
import copy
from ppocr.utils.logging import get_logger from ppocr.utils.logging import get_logger
...@@ -112,14 +113,14 @@ class BaseRecLabelEncode(object): ...@@ -112,14 +113,14 @@ class BaseRecLabelEncode(object):
dict_character = list(self.character_str) dict_character = list(self.character_str)
self.lower = True self.lower = True
else: else:
self.character_str = "" self.character_str = []
with open(character_dict_path, "rb") as fin: with open(character_dict_path, "rb") as fin:
lines = fin.readlines() lines = fin.readlines()
for line in lines: for line in lines:
line = line.decode('utf-8').strip("\n").strip("\r\n") line = line.decode('utf-8').strip("\n").strip("\r\n")
self.character_str += line self.character_str.append(line)
if use_space_char: if use_space_char:
self.character_str += " " self.character_str.append(" ")
dict_character = list(self.character_str) dict_character = list(self.character_str)
dict_character = self.add_special_char(dict_character) dict_character = self.add_special_char(dict_character)
self.dict = {} self.dict = {}
...@@ -1007,3 +1008,34 @@ class VQATokenLabelEncode(object): ...@@ -1007,3 +1008,34 @@ class VQATokenLabelEncode(object):
gt_label.extend([self.label2id_map[("i-" + label).upper()]] * gt_label.extend([self.label2id_map[("i-" + label).upper()]] *
(len(encode_res["input_ids"]) - 1)) (len(encode_res["input_ids"]) - 1))
return gt_label return gt_label
class MultiLabelEncode(BaseRecLabelEncode):
def __init__(self,
max_text_length,
character_dict_path=None,
use_space_char=False,
**kwargs):
super(MultiLabelEncode, self).__init__(
max_text_length, character_dict_path, use_space_char)
self.ctc_encode = CTCLabelEncode(max_text_length, character_dict_path,
use_space_char, **kwargs)
self.sar_encode = SARLabelEncode(max_text_length, character_dict_path,
use_space_char, **kwargs)
def __call__(self, data):
data_ctc = copy.deepcopy(data)
data_sar = copy.deepcopy(data)
data_out = dict()
data_out['img_path'] = data.get('img_path', None)
data_out['image'] = data['image']
ctc = self.ctc_encode.__call__(data_ctc)
sar = self.sar_encode.__call__(data_sar)
if ctc is None or sar is None:
return None
data_out['label_ctc'] = ctc['label']
data_out['label_sar'] = sar['label']
data_out['length'] = ctc['length']
return data_out
ppocr/data/imaug/rec_img_aug.py
View file @ 2b3f89f0
...@@ -16,6 +16,7 @@ import math ...@@ -16,6 +16,7 @@ import math
import cv2 import cv2
import numpy as np import numpy as np
import random import random
import copy
from PIL import Image from PIL import Image
from .text_image_aug import tia_perspective, tia_stretch, tia_distort from .text_image_aug import tia_perspective, tia_stretch, tia_distort
...@@ -32,13 +33,56 @@ class RecAug(object): ...@@ -32,13 +33,56 @@ class RecAug(object):
return data return data
class RecConAug(object):
def __init__(self,
prob=0.5,
image_shape=(32, 320, 3),
max_text_length=25,
ext_data_num=1,
**kwargs):
self.ext_data_num = ext_data_num
self.prob = prob
self.max_text_length = max_text_length
self.image_shape = image_shape
self.max_wh_ratio = self.image_shape[1] / self.image_shape[0]
def merge_ext_data(self, data, ext_data):
ori_w = round(data['image'].shape[1] / data['image'].shape[0] *
self.image_shape[0])
ext_w = round(ext_data['image'].shape[1] / ext_data['image'].shape[0] *
self.image_shape[0])
data['image'] = cv2.resize(data['image'], (ori_w, self.image_shape[0]))
ext_data['image'] = cv2.resize(ext_data['image'],
(ext_w, self.image_shape[0]))
data['image'] = np.concatenate(
[data['image'], ext_data['image']], axis=1)
data["label"] += ext_data["label"]
return data
def __call__(self, data):
rnd_num = random.random()
if rnd_num > self.prob:
return data
for idx, ext_data in enumerate(data["ext_data"]):
if len(data["label"]) + len(ext_data[
"label"]) > self.max_text_length:
break
concat_ratio = data['image'].shape[1] / data['image'].shape[
0] + ext_data['image'].shape[1] / ext_data['image'].shape[0]
if concat_ratio > self.max_wh_ratio:
break
data = self.merge_ext_data(data, ext_data)
data.pop("ext_data")
return data
class ClsResizeImg(object): class ClsResizeImg(object):
def __init__(self, image_shape, **kwargs): def __init__(self, image_shape, **kwargs):
self.image_shape = image_shape self.image_shape = image_shape
def __call__(self, data): def __call__(self, data):
img = data['image'] img = data['image']
norm_img = resize_norm_img(img, self.image_shape) norm_img, _ = resize_norm_img(img, self.image_shape)
data['image'] = norm_img data['image'] = norm_img
return data return data
...@@ -98,10 +142,13 @@ class RecResizeImg(object): ...@@ -98,10 +142,13 @@ class RecResizeImg(object):
def __call__(self, data): def __call__(self, data):
img = data['image'] img = data['image']
if self.infer_mode and self.character_dict_path is not None: if self.infer_mode and self.character_dict_path is not None:
norm_img = resize_norm_img_chinese(img, self.image_shape) norm_img, valid_ratio = resize_norm_img_chinese(img,
self.image_shape)
else: else:
norm_img = resize_norm_img(img, self.image_shape, self.padding) norm_img, valid_ratio = resize_norm_img(img, self.image_shape,
self.padding)
data['image'] = norm_img data['image'] = norm_img
data['valid_ratio'] = valid_ratio
return data return data
...@@ -160,6 +207,25 @@ class PRENResizeImg(object): ...@@ -160,6 +207,25 @@ class PRENResizeImg(object):
return data return data
class SVTRRecResizeImg(object):
def __init__(self,
image_shape,
infer_mode=False,
character_dict_path='./ppocr/utils/ppocr_keys_v1.txt',
padding=True,
**kwargs):
self.image_shape = image_shape
self.infer_mode = infer_mode
self.character_dict_path = character_dict_path
self.padding = padding
def __call__(self, data):
img = data['image']
norm_img = resize_norm_img_svtr(img, self.image_shape, self.padding)
data['image'] = norm_img
return data
def resize_norm_img_sar(img, image_shape, width_downsample_ratio=0.25): def resize_norm_img_sar(img, image_shape, width_downsample_ratio=0.25):
imgC, imgH, imgW_min, imgW_max = image_shape imgC, imgH, imgW_min, imgW_max = image_shape
h = img.shape[0] h = img.shape[0]
...@@ -220,7 +286,8 @@ def resize_norm_img(img, image_shape, padding=True): ...@@ -220,7 +286,8 @@ def resize_norm_img(img, image_shape, padding=True):
resized_image /= 0.5 resized_image /= 0.5
padding_im = np.zeros((imgC, imgH, imgW), dtype=np.float32) padding_im = np.zeros((imgC, imgH, imgW), dtype=np.float32)
padding_im[:, :, 0:resized_w] = resized_image padding_im[:, :, 0:resized_w] = resized_image
return padding_im valid_ratio = min(1.0, float(resized_w / imgW))
return padding_im, valid_ratio
def resize_norm_img_chinese(img, image_shape): def resize_norm_img_chinese(img, image_shape):
...@@ -230,7 +297,7 @@ def resize_norm_img_chinese(img, image_shape): ...@@ -230,7 +297,7 @@ def resize_norm_img_chinese(img, image_shape):
h, w = img.shape[0], img.shape[1] h, w = img.shape[0], img.shape[1]
ratio = w * 1.0 / h ratio = w * 1.0 / h
max_wh_ratio = max(max_wh_ratio, ratio) max_wh_ratio = max(max_wh_ratio, ratio)
imgW = int(32 * max_wh_ratio) imgW = int(imgH * max_wh_ratio)
if math.ceil(imgH * ratio) > imgW: if math.ceil(imgH * ratio) > imgW:
resized_w = imgW resized_w = imgW
else: else:
...@@ -246,7 +313,8 @@ def resize_norm_img_chinese(img, image_shape): ...@@ -246,7 +313,8 @@ def resize_norm_img_chinese(img, image_shape):
resized_image /= 0.5 resized_image /= 0.5
padding_im = np.zeros((imgC, imgH, imgW), dtype=np.float32) padding_im = np.zeros((imgC, imgH, imgW), dtype=np.float32)
padding_im[:, :, 0:resized_w] = resized_image padding_im[:, :, 0:resized_w] = resized_image
return padding_im valid_ratio = min(1.0, float(resized_w / imgW))
return padding_im, valid_ratio
def resize_norm_img_srn(img, image_shape): def resize_norm_img_srn(img, image_shape):
...@@ -276,6 +344,58 @@ def resize_norm_img_srn(img, image_shape): ...@@ -276,6 +344,58 @@ def resize_norm_img_srn(img, image_shape):
return np.reshape(img_black, (c, row, col)).astype(np.float32) return np.reshape(img_black, (c, row, col)).astype(np.float32)
def resize_norm_img_svtr(img, image_shape, padding=False):
imgC, imgH, imgW = image_shape
h = img.shape[0]
w = img.shape[1]
if not padding:
if h > 2.0 * w:
image = Image.fromarray(img)
image1 = image.rotate(90, expand=True)
image2 = image.rotate(-90, expand=True)
img1 = np.array(image1)
img2 = np.array(image2)
else:
img1 = copy.deepcopy(img)
img2 = copy.deepcopy(img)
resized_image = cv2.resize(
img, (imgW, imgH), interpolation=cv2.INTER_LINEAR)
resized_image1 = cv2.resize(
img1, (imgW, imgH), interpolation=cv2.INTER_LINEAR)
resized_image2 = cv2.resize(
img2, (imgW, imgH), interpolation=cv2.INTER_LINEAR)
resized_w = imgW
else:
ratio = w / float(h)
if math.ceil(imgH * ratio) > imgW:
resized_w = imgW
else:
resized_w = int(math.ceil(imgH * ratio))
resized_image = cv2.resize(img, (resized_w, imgH))
resized_image = resized_image.astype('float32')
resized_image1 = resized_image1.astype('float32')
resized_image2 = resized_image2.astype('float32')
if image_shape[0] == 1:
resized_image = resized_image / 255
resized_image = resized_image[np.newaxis, :]
else:
resized_image = resized_image.transpose((2, 0, 1)) / 255
resized_image1 = resized_image1.transpose((2, 0, 1)) / 255
resized_image2 = resized_image2.transpose((2, 0, 1)) / 255
resized_image -= 0.5
resized_image /= 0.5
resized_image1 -= 0.5
resized_image1 /= 0.5
resized_image2 -= 0.5
resized_image2 /= 0.5
padding_im = np.zeros((3, imgC, imgH, imgW), dtype=np.float32)
padding_im[0, :, :, 0:resized_w] = resized_image
padding_im[1, :, :, 0:resized_w] = resized_image1
padding_im[2, :, :, 0:resized_w] = resized_image2
return padding_im
def srn_other_inputs(image_shape, num_heads, max_text_length): def srn_other_inputs(image_shape, num_heads, max_text_length):
imgC, imgH, imgW = image_shape imgC, imgH, imgW = image_shape
......
ppocr/data/simple_dataset.py
View file @ 2b3f89f0
...@@ -49,7 +49,8 @@ class SimpleDataSet(Dataset): ...@@ -49,7 +49,8 @@ class SimpleDataSet(Dataset):
if self.mode == "train" and self.do_shuffle: if self.mode == "train" and self.do_shuffle:
self.shuffle_data_random() self.shuffle_data_random()
self.ops = create_operators(dataset_config['transforms'], global_config) self.ops = create_operators(dataset_config['transforms'], global_config)
self.ext_op_transform_idx = dataset_config.get("ext_op_transform_idx",
2)
self.need_reset = True in [x < 1 for x in ratio_list] self.need_reset = True in [x < 1 for x in ratio_list]
def get_image_info_list(self, file_list, ratio_list): def get_image_info_list(self, file_list, ratio_list):
...@@ -87,7 +88,7 @@ class SimpleDataSet(Dataset): ...@@ -87,7 +88,7 @@ class SimpleDataSet(Dataset):
if hasattr(op, 'ext_data_num'): if hasattr(op, 'ext_data_num'):
ext_data_num = getattr(op, 'ext_data_num') ext_data_num = getattr(op, 'ext_data_num')
break break
load_data_ops = self.ops[:2] load_data_ops = self.ops[:self.ext_op_transform_idx]
ext_data = [] ext_data = []
while len(ext_data) < ext_data_num: while len(ext_data) < ext_data_num:
...@@ -108,8 +109,11 @@ class SimpleDataSet(Dataset): ...@@ -108,8 +109,11 @@ class SimpleDataSet(Dataset):
data['image'] = img data['image'] = img
data = transform(data, load_data_ops) data = transform(data, load_data_ops)
if data is None or data['polys'].shape[1] != 4: if data is None:
continue continue
if 'polys' in data.keys():
if data['polys'].shape[1] != 4:
continue
ext_data.append(data) ext_data.append(data)
return ext_data return ext_data
......
ppocr/losses/__init__.py
View file @ 2b3f89f0
...@@ -34,6 +34,7 @@ from .rec_nrtr_loss import NRTRLoss ...@@ -34,6 +34,7 @@ from .rec_nrtr_loss import NRTRLoss
from .rec_sar_loss import SARLoss from .rec_sar_loss import SARLoss
from .rec_aster_loss import AsterLoss from .rec_aster_loss import AsterLoss
from .rec_pren_loss import PRENLoss from .rec_pren_loss import PRENLoss
from .rec_multi_loss import MultiLoss
# cls loss # cls loss
from .cls_loss import ClsLoss from .cls_loss import ClsLoss
...@@ -60,7 +61,7 @@ def build_loss(config): ...@@ -60,7 +61,7 @@ def build_loss(config):
'DBLoss', 'PSELoss', 'EASTLoss', 'SASTLoss', 'FCELoss', 'CTCLoss', 'DBLoss', 'PSELoss', 'EASTLoss', 'SASTLoss', 'FCELoss', 'CTCLoss',
'ClsLoss', 'AttentionLoss', 'SRNLoss', 'PGLoss', 'CombinedLoss', 'ClsLoss', 'AttentionLoss', 'SRNLoss', 'PGLoss', 'CombinedLoss',
'NRTRLoss', 'TableAttentionLoss', 'SARLoss', 'AsterLoss', 'SDMGRLoss', 'NRTRLoss', 'TableAttentionLoss', 'SARLoss', 'AsterLoss', 'SDMGRLoss',
'VQASerTokenLayoutLMLoss', 'LossFromOutput', 'PRENLoss' 'VQASerTokenLayoutLMLoss', 'LossFromOutput', 'PRENLoss', 'MultiLoss'
] ]
config = copy.deepcopy(config) config = copy.deepcopy(config)
module_name = config.pop('name') module_name = config.pop('name')
......
ppocr/losses/basic_loss.py
View file @ 2b3f89f0
...@@ -106,8 +106,8 @@ class DMLLoss(nn.Layer): ...@@ -106,8 +106,8 @@ class DMLLoss(nn.Layer):
def forward(self, out1, out2): def forward(self, out1, out2):
if self.act is not None: if self.act is not None:
out1 = self.act(out1) out1 = self.act(out1) + 1e-10
out2 = self.act(out2) out2 = self.act(out2) + 1e-10
if self.use_log: if self.use_log:
# for recognition distillation, log is needed for feature map # for recognition distillation, log is needed for feature map
log_out1 = paddle.log(out1) log_out1 = paddle.log(out1)
......
ppocr/losses/combined_loss.py
View file @ 2b3f89f0
...@@ -18,8 +18,10 @@ import paddle.nn as nn ...@@ -18,8 +18,10 @@ import paddle.nn as nn
from .rec_ctc_loss import CTCLoss from .rec_ctc_loss import CTCLoss
from .center_loss import CenterLoss from .center_loss import CenterLoss
from .ace_loss import ACELoss from .ace_loss import ACELoss
from .rec_sar_loss import SARLoss
from .distillation_loss import DistillationCTCLoss from .distillation_loss import DistillationCTCLoss
from .distillation_loss import DistillationSARLoss
from .distillation_loss import DistillationDMLLoss from .distillation_loss import DistillationDMLLoss
from .distillation_loss import DistillationDistanceLoss, DistillationDBLoss, DistillationDilaDBLoss from .distillation_loss import DistillationDistanceLoss, DistillationDBLoss, DistillationDilaDBLoss
......
ppocr/losses/distillation_loss.py
View file @ 2b3f89f0
...@@ -18,6 +18,7 @@ import numpy as np ...@@ -18,6 +18,7 @@ import numpy as np
import cv2 import cv2
from .rec_ctc_loss import CTCLoss from .rec_ctc_loss import CTCLoss
from .rec_sar_loss import SARLoss
from .basic_loss import DMLLoss from .basic_loss import DMLLoss
from .basic_loss import DistanceLoss from .basic_loss import DistanceLoss
from .det_db_loss import DBLoss from .det_db_loss import DBLoss
...@@ -46,11 +47,15 @@ class DistillationDMLLoss(DMLLoss): ...@@ -46,11 +47,15 @@ class DistillationDMLLoss(DMLLoss):
act=None, act=None,
use_log=False, use_log=False,
key=None, key=None,
multi_head=False,
dis_head='ctc',
maps_name=None, maps_name=None,
name="dml"): name="dml"):
super().__init__(act=act, use_log=use_log) super().__init__(act=act, use_log=use_log)
assert isinstance(model_name_pairs, list) assert isinstance(model_name_pairs, list)
self.key = key self.key = key
self.multi_head = multi_head
self.dis_head = dis_head
self.model_name_pairs = self._check_model_name_pairs(model_name_pairs) self.model_name_pairs = self._check_model_name_pairs(model_name_pairs)
self.name = name self.name = name
self.maps_name = self._check_maps_name(maps_name) self.maps_name = self._check_maps_name(maps_name)
...@@ -97,7 +102,11 @@ class DistillationDMLLoss(DMLLoss): ...@@ -97,7 +102,11 @@ class DistillationDMLLoss(DMLLoss):
out2 = out2[self.key] out2 = out2[self.key]
if self.maps_name is None: if self.maps_name is None:
loss = super().forward(out1, out2) if self.multi_head:
loss = super().forward(out1[self.dis_head],
out2[self.dis_head])
else:
loss = super().forward(out1, out2)
if isinstance(loss, dict): if isinstance(loss, dict):
for key in loss: for key in loss:
loss_dict["{}_{}_{}_{}".format(key, pair[0], pair[1], loss_dict["{}_{}_{}_{}".format(key, pair[0], pair[1],
...@@ -123,11 +132,16 @@ class DistillationDMLLoss(DMLLoss): ...@@ -123,11 +132,16 @@ class DistillationDMLLoss(DMLLoss):
class DistillationCTCLoss(CTCLoss): class DistillationCTCLoss(CTCLoss):
def __init__(self, model_name_list=[], key=None, name="loss_ctc"): def __init__(self,
model_name_list=[],
key=None,
multi_head=False,
name="loss_ctc"):
super().__init__() super().__init__()
self.model_name_list = model_name_list self.model_name_list = model_name_list
self.key = key self.key = key
self.name = name self.name = name
self.multi_head = multi_head
def forward(self, predicts, batch): def forward(self, predicts, batch):
loss_dict = dict() loss_dict = dict()
...@@ -135,7 +149,45 @@ class DistillationCTCLoss(CTCLoss): ...@@ -135,7 +149,45 @@ class DistillationCTCLoss(CTCLoss):
out = predicts[model_name] out = predicts[model_name]
if self.key is not None: if self.key is not None:
out = out[self.key] out = out[self.key]
loss = super().forward(out, batch) if self.multi_head:
assert 'ctc' in out, 'multi head has multi out'
loss = super().forward(out['ctc'], batch[:2] + batch[3:])
else:
loss = super().forward(out, batch)
if isinstance(loss, dict):
for key in loss:
loss_dict["{}_{}_{}".format(self.name, model_name,
idx)] = loss[key]
else:
loss_dict["{}_{}".format(self.name, model_name)] = loss
return loss_dict
class DistillationSARLoss(SARLoss):
def __init__(self,
model_name_list=[],
key=None,
multi_head=False,
name="loss_sar",
**kwargs):
ignore_index = kwargs.get('ignore_index', 92)
super().__init__(ignore_index=ignore_index)
self.model_name_list = model_name_list
self.key = key
self.name = name
self.multi_head = multi_head
def forward(self, predicts, batch):
loss_dict = dict()
for idx, model_name in enumerate(self.model_name_list):
out = predicts[model_name]
if self.key is not None:
out = out[self.key]
if self.multi_head:
assert 'sar' in out, 'multi head has multi out'
loss = super().forward(out['sar'], batch[:1] + batch[2:])
else:
loss = super().forward(out, batch)
if isinstance(loss, dict): if isinstance(loss, dict):
for key in loss: for key in loss:
loss_dict["{}_{}_{}".format(self.name, model_name, loss_dict["{}_{}_{}".format(self.name, model_name,
......
ppocr/losses/rec_multi_loss.py 0 → 100644
View file @ 2b3f89f0
# copyright (c) 2022 PaddlePaddle Authors. All Rights Reserve.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import paddle
from paddle import nn
from .rec_ctc_loss import CTCLoss
from .rec_sar_loss import SARLoss
class MultiLoss(nn.Layer):
def __init__(self, **kwargs):
super().__init__()
self.loss_funcs = {}
self.loss_list = kwargs.pop('loss_config_list')
self.weight_1 = kwargs.get('weight_1', 1.0)
self.weight_2 = kwargs.get('weight_2', 1.0)
self.gtc_loss = kwargs.get('gtc_loss', 'sar')
for loss_info in self.loss_list:
for name, param in loss_info.items():
if param is not None:
kwargs.update(param)
loss = eval(name)(**kwargs)
self.loss_funcs[name] = loss
def forward(self, predicts, batch):
self.total_loss = {}
total_loss = 0.0
# batch [image, label_ctc, label_sar, length, valid_ratio]
for name, loss_func in self.loss_funcs.items():
if name == 'CTCLoss':
loss = loss_func(predicts['ctc'],
batch[:2] + batch[3:])['loss'] * self.weight_1
elif name == 'SARLoss':
loss = loss_func(predicts['sar'],
batch[:1] + batch[2:])['loss'] * self.weight_2
else:
raise NotImplementedError(
'{} is not supported in MultiLoss yet'.format(name))
self.total_loss[name] = loss
total_loss += loss
self.total_loss['loss'] = total_loss
return self.total_loss
ppocr/losses/rec_sar_loss.py
View file @ 2b3f89f0
...@@ -9,8 +9,9 @@ from paddle import nn ...@@ -9,8 +9,9 @@ from paddle import nn
class SARLoss(nn.Layer): class SARLoss(nn.Layer):
def __init__(self, **kwargs): def __init__(self, **kwargs):
super(SARLoss, self).__init__() super(SARLoss, self).__init__()
ignore_index = kwargs.get('ignore_index', 92) # 6626
self.loss_func = paddle.nn.loss.CrossEntropyLoss( self.loss_func = paddle.nn.loss.CrossEntropyLoss(
reduction="mean", ignore_index=92) reduction="mean", ignore_index=ignore_index)
def forward(self, predicts, batch): def forward(self, predicts, batch):
predict = predicts[:, : predict = predicts[:, :
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment