init

deploy/slim/prune/README.md

@@ -0,0 +1,67 @@
+
+# PP-OCR模型裁剪
+
+复杂的模型有利于提高模型的性能，但也导致模型中存在一定冗余，模型裁剪通过移出网络模型中的子模型来减少这种冗余，达到减少模型计算复杂度，提高模型推理性能的目的。
+本教程将介绍如何使用飞桨模型压缩库PaddleSlim做PaddleOCR模型的压缩。
+[PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim)集成了模型剪枝、量化（包括量化训练和离线量化）、蒸馏和神经网络搜索等多种业界常用且领先的模型压缩功能，如果您感兴趣，可以关注并了解。
+
+
+在开始本教程之前，建议先了解：
+1. [PaddleOCR模型的训练方法](../../../doc/doc_ch/training.md)
+2. [模型裁剪教程](https://github.com/PaddlePaddle/PaddleSlim/blob/release%2F2.0.0/docs/zh_cn/tutorials/pruning/dygraph/filter_pruning.md)
+
+## 快速开始
+
+模型裁剪主要包括四个步骤：
+
+1. 安装 PaddleSlim
+2. 准备训练好的模型
+3. 敏感度分析、裁剪训练
+4. 导出模型、预测部署
+
+### 1. 安装PaddleSlim
+
+```bash
+git clone https://github.com/PaddlePaddle/PaddleSlim.git
+cd PaddleSlim
+git checkout develop
+python3 setup.py install
+```
+
+### 2. 获取预训练模型
+模型裁剪需要加载事先训练好的模型，PaddleOCR也提供了一系列[模型](../../../doc/doc_ch/models_list.md)，开发者可根据需要自行选择模型或使用自己的模型。
+
+### 3. 敏感度分析训练
+
+加载预训练模型后，通过对现有模型的每个网络层进行敏感度分析，得到敏感度文件：sen.pickle，可以通过PaddleSlim提供的[接口](https://github.com/PaddlePaddle/PaddleSlim/blob/9b01b195f0c4bc34a1ab434751cb260e13d64d9e/paddleslim/dygraph/prune/filter_pruner.py#L75)加载文件，获得各网络层在不同裁剪比例下的精度损失。从而了解各网络层冗余度，决定每个网络层的裁剪比例。
+敏感度文件内容格式：
+```
+sen.pickle(Dict){
+            'layer_weight_name_0': sens_of_each_ratio(Dict){'pruning_ratio_0': acc_loss, 'pruning_ratio_1': acc_loss}
+            'layer_weight_name_1': sens_of_each_ratio(Dict){'pruning_ratio_0': acc_loss, 'pruning_ratio_1': acc_loss}
+        }
+
+例子：
+        {
+            'conv10_expand_weights': {0.1: 0.006509952684312718, 0.2: 0.01827734339798862, 0.3: 0.014528405644659832, 0.6: 0.06536008804270439, 0.8: 0.11798612250664964, 0.7: 0.12391408417493704, 0.4: 0.030615754498018757, 0.5: 0.047105205602406594}
+            'conv10_linear_weights': {0.1: 0.05113190831455035, 0.2: 0.07705573833558801, 0.3: 0.12096721757739311, 0.6: 0.5135061352930738, 0.8: 0.7908166677143281, 0.7: 0.7272187676899062, 0.4: 0.1819252083008504, 0.5: 0.3728054727792405}
+        }
+```
+
+加载敏感度文件后会返回一个字典，字典中的keys为网络模型参数模型的名字，values为一个字典，里面保存了相应网络层的裁剪敏感度信息。例如在例子中，conv10_expand_weights所对应的网络层在裁掉10%的卷积核后模型性能相较原模型会下降0.65%，详细信息可见[PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim/blob/develop/docs/zh_cn/algo/algo.md#2-%E5%8D%B7%E7%A7%AF%E6%A0%B8%E5%89%AA%E8%A3%81%E5%8E%9F%E7%90%86)
+
+进入PaddleOCR根目录，通过以下命令对模型进行敏感度分析训练：
+```bash
+python3.7 deploy/slim/prune/sensitivity_anal.py -c configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml -o Global.pretrained_model="your trained model" Global.save_model_dir=./output/prune_model/
+```
+
+### 4. 导出模型、预测部署
+
+在得到裁剪训练保存的模型后，我们可以将其导出为inference_model：
+```bash
+pytho3.7 deploy/slim/prune/export_prune_model.py -c configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml -o Global.pretrained_model=./output/det_db/best_accuracy  Global.save_inference_dir=./prune/prune_inference_model
+```
+
+inference model的预测和部署参考：
+1. [inference model python端预测](../../../doc/doc_ch/inference.md)
+2. [inference model C++预测](../../cpp_infer/readme.md)

deploy/slim/prune/README_en.md

@@ -0,0 +1,73 @@
+
+# PP-OCR Models Pruning
+
+Generally, a more complex model would achieve better performance in the task, but it also leads to some redundancy in the model. Model Pruning is a technique that reduces this redundancy by removing the sub-models in the neural network model, so as to reduce model calculation complexity and improve model inference performance.
+
+This example uses PaddleSlim provided[APIs of Pruning](https://github.com/PaddlePaddle/PaddleSlim/tree/develop/docs/zh_cn/api_cn/dygraph/pruners) to compress the OCR model.
+[PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim), an open source library which integrates model pruning, quantization (including quantization training and offline quantization), distillation, neural network architecture search, and many other commonly used and leading model compression technique in the industry.
+
+It is recommended that you could understand following pages before reading this example：
+1. [PaddleOCR training methods](../../../doc/doc_ch/quickstart.md)
+2. [The demo of prune](https://github.com/PaddlePaddle/PaddleSlim/blob/release%2F2.0.0/docs/zh_cn/tutorials/pruning/dygraph/filter_pruning.md)
+
+## Quick start
+
+Five steps for OCR model prune:
+1. Install PaddleSlim
+2. Prepare the trained model
+3. Sensitivity analysis and tailoring training
+4. Export model, predict deployment
+
+### 1. Install PaddleSlim
+
+```bash
+git clone https://github.com/PaddlePaddle/PaddleSlim.git
+cd PaddleSlim
+git checkout develop
+python3 setup.py install
+```
+
+
+### 2. Download Pre-trained Model
+Model prune needs to load pre-trained models.
+PaddleOCR also provides a series of [models](../../../doc/doc_en/models_list_en.md). Developers can choose their own models or use their own models according to their needs.
+
+
+### 3. Pruning sensitivity analysis
+
+  After the pre-trained model is loaded, sensitivity analysis is performed on each network layer of the model to understand the redundancy of each network layer, and save a sensitivity file which named: sen.pickle.  After that, user could load the sensitivity file via the [methods provided by PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim/blob/develop/paddleslim/prune/sensitive.py#L221) and determining the pruning ratio of each network layer automatically. For specific details of sensitivity analysis, see：[Sensitivity analysis](https://github.com/PaddlePaddle/PaddleSlim/blob/develop/docs/en/tutorials/image_classification_sensitivity_analysis_tutorial_en.md)
+  The data format of sensitivity file：
+
+```      
+sen.pickle(Dict){
+              'layer_weight_name_0': sens_of_each_ratio(Dict){'pruning_ratio_0': acc_loss, 'pruning_ratio_1': acc_loss}
+              'layer_weight_name_1': sens_of_each_ratio(Dict){'pruning_ratio_0': acc_loss, 'pruning_ratio_1': acc_loss}
+          }
+example：
+          {
+              'conv10_expand_weights': {0.1: 0.006509952684312718, 0.2: 0.01827734339798862, 0.3: 0.014528405644659832, 0.6: 0.06536008804270439, 0.8: 0.11798612250664964, 0.7: 0.12391408417493704, 0.4: 0.030615754498018757, 0.5: 0.047105205602406594}
+              'conv10_linear_weights': {0.1: 0.05113190831455035, 0.2: 0.07705573833558801, 0.3: 0.12096721757739311, 0.6: 0.5135061352930738, 0.8: 0.7908166677143281, 0.7: 0.7272187676899062, 0.4: 0.1819252083008504, 0.5: 0.3728054727792405}
+          }
+  The function would return a dict after loading the sensitivity file. The keys of the dict are name of parameters in each layer. And the value of key is the information about pruning sensitivity of corresponding layer. In example, pruning 10% filter of the layer corresponding to conv10_expand_weights would lead to 0.65% degradation of model performance. The details could be seen at: [Sensitivity analysis](https://github.com/PaddlePaddle/PaddleSlim/blob/release/2.0-alpha/docs/zh_cn/algo/algo.md)
+```
+
+  The function would return a dict after loading the sensitivity file. The keys of the dict are name of parameters in each layer. And the value of key is the information about pruning sensitivity of corresponding layer. In example, pruning 10% filter of the layer corresponding to conv10_expand_weights would lead to 0.65% degradation of model performance. The details could be seen at: [Sensitivity analysis](https://github.com/PaddlePaddle/PaddleSlim/blob/develop/docs/zh_cn/algo/algo.md#2-%E5%8D%B7%E7%A7%AF%E6%A0%B8%E5%89%AA%E8%A3%81%E5%8E%9F%E7%90%86)
+
+
+Enter the PaddleOCR root directory，perform sensitivity analysis on the model with the following command：
+
+```bash
+python3.7 deploy/slim/prune/sensitivity_anal.py -c configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml -o Global.pretrained_model="your trained model"  Global.save_model_dir=./output/prune_model/
+```
+
+
+### 5.  Export inference model and deploy it
+
+We can export the pruned model as inference_model for deployment:
+```bash
+python deploy/slim/prune/export_prune_model.py -c configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml  -o Global.pretrained_model=./output/det_db/best_accuracy  Global.save_inference_dir=./prune/prune_inference_model
+```
+
+Reference for prediction and deployment of inference model:
+1. [inference model python prediction](../../../doc/doc_en/inference_en.md)
+2. [inference model C++ prediction](../../cpp_infer/readme_en.md)

deploy/slim/prune/export_prune_model.py

@@ -0,0 +1,140 @@
+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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 os
+import sys
+
+__dir__ = os.path.dirname(__file__)
+sys.path.append(__dir__)
+sys.path.append(os.path.join(__dir__, '..', '..', '..'))
+sys.path.append(os.path.join(__dir__, '..', '..', '..', 'tools'))
+
+import paddle
+from ppocr.data import build_dataloader, set_signal_handlers
+from ppocr.modeling.architectures import build_model
+
+from ppocr.postprocess import build_post_process
+from ppocr.metrics import build_metric
+from ppocr.utils.save_load import load_model
+import tools.program as program
+
+
+def main(config, device, logger, vdl_writer):
+
+    global_config = config['Global']
+
+    # build dataloader
+    set_signal_handlers()
+    valid_dataloader = build_dataloader(config, 'Eval', device, logger)
+
+    # build post process
+    post_process_class = build_post_process(config['PostProcess'],
+                                            global_config)
+
+    # build model
+    # for rec algorithm
+    if hasattr(post_process_class, 'character'):
+        char_num = len(getattr(post_process_class, 'character'))
+        config['Architecture']["Head"]['out_channels'] = char_num
+    model = build_model(config['Architecture'])
+
+    if config['Architecture']['model_type'] == 'det':
+        input_shape = [1, 3, 640, 640]
+    elif config['Architecture']['model_type'] == 'rec':
+        input_shape = [1, 3, 32, 320]
+
+    flops = paddle.flops(model, input_shape)
+    logger.info("FLOPs before pruning: {}".format(flops))
+
+    from paddleslim.dygraph import FPGMFilterPruner
+    model.train()
+    pruner = FPGMFilterPruner(model, input_shape)
+
+    # build metric
+    eval_class = build_metric(config['Metric'])
+
+    def eval_fn():
+        metric = program.eval(model, valid_dataloader, post_process_class,
+                              eval_class)
+        if config['Architecture']['model_type'] == 'det':
+            main_indicator = 'hmean'
+        else:
+            main_indicator = 'acc'
+        logger.info("metric[{}]: {}".format(main_indicator, metric[
+            main_indicator]))
+        return metric[main_indicator]
+
+    params_sensitive = pruner.sensitive(
+        eval_func=eval_fn,
+        sen_file="./sen.pickle",
+        skip_vars=[
+            "conv2d_57.w_0", "conv2d_transpose_2.w_0", "conv2d_transpose_3.w_0"
+        ])
+
+    logger.info(
+        "The sensitivity analysis results of model parameters saved in sen.pickle"
+    )
+    # calculate pruned params's ratio
+    params_sensitive = pruner._get_ratios_by_loss(params_sensitive, loss=0.02)
+    for key in params_sensitive.keys():
+        logger.info("{}, {}".format(key, params_sensitive[key]))
+
+    plan = pruner.prune_vars(params_sensitive, [0])
+
+    flops = paddle.flops(model, input_shape)
+    logger.info("FLOPs after pruning: {}".format(flops))
+
+    # load pretrain model
+    load_model(config, model)
+    metric = program.eval(model, valid_dataloader, post_process_class,
+                          eval_class)
+    if config['Architecture']['model_type'] == 'det':
+        main_indicator = 'hmean'
+    else:
+        main_indicator = 'acc'
+    logger.info("metric['']: {}".format(main_indicator, metric[main_indicator]))
+
+    # start export model
+    from paddle.jit import to_static
+
+    infer_shape = [3, -1, -1]
+    if config['Architecture']['model_type'] == "rec":
+        infer_shape = [3, 32, -1]  # for rec model, H must be 32
+
+        if 'Transform' in config['Architecture'] and config['Architecture'][
+                'Transform'] is not None and config['Architecture'][
+                    'Transform']['name'] == 'TPS':
+            logger.info(
+                'When there is tps in the network, variable length input is not supported, and the input size needs to be the same as during training'
+            )
+            infer_shape[-1] = 100
+    model = to_static(
+        model,
+        input_spec=[
+            paddle.static.InputSpec(
+                shape=[None] + infer_shape, dtype='float32')
+        ])
+
+    save_path = '{}/inference'.format(config['Global']['save_inference_dir'])
+    paddle.jit.save(model, save_path)
+    logger.info('inference model is saved to {}'.format(save_path))
+
+
+if __name__ == '__main__':
+    config, device, logger, vdl_writer = program.preprocess(is_train=True)
+    main(config, device, logger, vdl_writer)

deploy/slim/prune/sensitivity_anal.py

@@ -0,0 +1,176 @@
+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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 os
+import sys
+
+__dir__ = os.path.dirname(__file__)
+sys.path.append(__dir__)
+sys.path.append(os.path.join(__dir__, '..', '..', '..'))
+sys.path.append(os.path.join(__dir__, '..', '..', '..', 'tools'))
+
+import paddle
+import paddle.distributed as dist
+from ppocr.data import build_dataloader, set_signal_handlers
+from ppocr.modeling.architectures import build_model
+from ppocr.losses import build_loss
+from ppocr.optimizer import build_optimizer
+from ppocr.postprocess import build_post_process
+from ppocr.metrics import build_metric
+from ppocr.utils.save_load import load_model
+import tools.program as program
+
+dist.get_world_size()
+
+
+def get_pruned_params(parameters):
+    params = []
+
+    for param in parameters:
+        if len(
+                param.shape
+        ) == 4 and 'depthwise' not in param.name and 'transpose' not in param.name and "conv2d_57" not in param.name and "conv2d_56" not in param.name:
+            params.append(param.name)
+    return params
+
+
+def main(config, device, logger, vdl_writer):
+    # init dist environment
+    if config['Global']['distributed']:
+        dist.init_parallel_env()
+
+    global_config = config['Global']
+
+    # build dataloader
+    set_signal_handlers()
+    train_dataloader = build_dataloader(config, 'Train', device, logger)
+    if config['Eval']:
+        valid_dataloader = build_dataloader(config, 'Eval', device, logger)
+    else:
+        valid_dataloader = None
+
+    # build post process
+    post_process_class = build_post_process(config['PostProcess'],
+                                            global_config)
+
+    # build model
+    # for rec algorithm
+    if hasattr(post_process_class, 'character'):
+        char_num = len(getattr(post_process_class, 'character'))
+        config['Architecture']["Head"]['out_channels'] = char_num
+    model = build_model(config['Architecture'])
+    if config['Architecture']['model_type'] == 'det':
+        input_shape = [1, 3, 640, 640]
+    elif config['Architecture']['model_type'] == 'rec':
+        input_shape = [1, 3, 32, 320]
+    flops = paddle.flops(model, input_shape)
+
+    logger.info("FLOPs before pruning: {}".format(flops))
+
+    from paddleslim.dygraph import FPGMFilterPruner
+    model.train()
+
+    pruner = FPGMFilterPruner(model, input_shape)
+
+    # build loss
+    loss_class = build_loss(config['Loss'])
+
+    # build optim
+    optimizer, lr_scheduler = build_optimizer(
+        config['Optimizer'],
+        epochs=config['Global']['epoch_num'],
+        step_each_epoch=len(train_dataloader),
+        model=model)
+
+    # build metric
+    eval_class = build_metric(config['Metric'])
+    # load pretrain model
+    pre_best_model_dict = load_model(config, model, optimizer)
+
+    logger.info('train dataloader has {} iters, valid dataloader has {} iters'.
+                format(len(train_dataloader), len(valid_dataloader)))
+    # build metric
+    eval_class = build_metric(config['Metric'])
+
+    logger.info('train dataloader has {} iters, valid dataloader has {} iters'.
+                format(len(train_dataloader), len(valid_dataloader)))
+
+    def eval_fn():
+        metric = program.eval(model, valid_dataloader, post_process_class,
+                              eval_class, False)
+        if config['Architecture']['model_type'] == 'det':
+            main_indicator = 'hmean'
+        else:
+            main_indicator = 'acc'
+
+        logger.info("metric[{}]: {}".format(main_indicator, metric[
+            main_indicator]))
+        return metric[main_indicator]
+
+    run_sensitive_analysis = False
+    """
+    run_sensitive_analysis=True: 
+        Automatically compute the sensitivities of convolutions in a model. 
+        The sensitivity of a convolution is the losses of accuracy on test dataset in 
+        differenct pruned ratios. The sensitivities can be used to get a group of best 
+        ratios with some condition.
+    
+    run_sensitive_analysis=False: 
+        Set prune trim ratio to a fixed value, such as 10%. The larger the value, 
+        the more convolution weights will be cropped.
+
+    """
+
+    if run_sensitive_analysis:
+        params_sensitive = pruner.sensitive(
+            eval_func=eval_fn,
+            sen_file="./deploy/slim/prune/sen.pickle",
+            skip_vars=[
+                "conv2d_57.w_0", "conv2d_transpose_2.w_0",
+                "conv2d_transpose_3.w_0"
+            ])
+        logger.info(
+            "The sensitivity analysis results of model parameters saved in sen.pickle"
+        )
+        # calculate pruned params's ratio
+        params_sensitive = pruner._get_ratios_by_loss(
+            params_sensitive, loss=0.02)
+        for key in params_sensitive.keys():
+            logger.info("{}, {}".format(key, params_sensitive[key]))
+    else:
+        params_sensitive = {}
+        for param in model.parameters():
+            if 'transpose' not in param.name and 'linear' not in param.name:
+                # set prune ratio as 10%. The larger the value, the more convolution weights will be cropped
+                params_sensitive[param.name] = 0.1
+
+    plan = pruner.prune_vars(params_sensitive, [0])
+
+    flops = paddle.flops(model, input_shape)
+    logger.info("FLOPs after pruning: {}".format(flops))
+
+    # start train
+
+    program.train(config, train_dataloader, valid_dataloader, device, model,
+                  loss_class, optimizer, lr_scheduler, post_process_class,
+                  eval_class, pre_best_model_dict, logger, vdl_writer)
+
+
+if __name__ == '__main__':
+    config, device, logger, vdl_writer = program.preprocess(is_train=True)
+    main(config, device, logger, vdl_writer)

deploy/slim/quantization/README.md

@@ -0,0 +1,60 @@
+
+# PP-OCR模型量化
+复杂的模型有利于提高模型的性能，但也导致模型中存在一定冗余，模型量化将全精度缩减到定点数减少这种冗余，达到减少模型计算复杂度，提高模型推理性能的目的。
+模型量化可以在基本不损失模型的精度的情况下，将FP32精度的模型参数转换为Int8精度，减小模型参数大小并加速计算，使用量化后的模型在移动端等部署时更具备速度优势。
+
+本教程将介绍如何使用飞桨模型压缩库PaddleSlim做PaddleOCR模型的压缩。
+[PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim) 集成了模型剪枝、量化（包括量化训练和离线量化）、蒸馏和神经网络搜索等多种业界常用且领先的模型压缩功能，如果您感兴趣，可以关注并了解。
+
+在开始本教程之前，建议先了解[PaddleOCR模型的训练方法](../../../doc/doc_ch/training.md)以及[PaddleSlim](https://paddleslim.readthedocs.io/zh_CN/latest/index.html)
+
+
+## 快速开始
+量化多适用于轻量模型在移动端的部署，当训练出一个模型后，如果希望进一步的压缩模型大小并加速预测，可使用量化的方法压缩模型。
+
+模型量化主要包括五个步骤：
+1. 安装 PaddleSlim
+2. 准备训练好的模型
+3. 量化训练
+4. 导出量化推理模型
+5. 量化模型预测部署
+
+### 1. 安装PaddleSlim
+
+```bash
+pip3 install paddleslim==2.3.2
+```
+
+### 2. 准备训练好的模型
+
+PaddleOCR提供了一系列训练好的[模型](../../../doc/doc_ch/models_list.md)，如果待量化的模型不在列表中，需要按照[常规训练](../../../doc/doc_ch/quickstart.md)方法得到训练好的模型。
+
+### 3. 量化训练
+量化训练包括离线量化训练和在线量化训练，在线量化训练效果更好，需加载预训练模型，在定义好量化策略后即可对模型进行量化。
+
+
+量化训练的代码位于slim/quantization/quant.py 中，比如训练检测模型，以PPOCRv3检测模型为例，训练指令如下：
+```
+# 下载检测预训练模型：
+wget https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_distill_train.tar
+tar xf ch_PP-OCRv3_det_distill_train.tar
+
+python deploy/slim/quantization/quant.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o Global.pretrained_model='./ch_PP-OCRv3_det_distill_train/best_accuracy'   Global.save_model_dir=./output/quant_model_distill/
+```
+如果要训练识别模型的量化，修改配置文件和加载的模型参数即可。
+
+### 4. 导出模型
+
+在得到量化训练保存的模型后，我们可以将其导出为inference_model，用于预测部署：
+
+```bash
+python deploy/slim/quantization/export_model.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o Global.checkpoints=output/quant_model/best_accuracy Global.save_inference_dir=./output/quant_inference_model
+```
+
+### 5. 量化模型部署
+
+上述步骤导出的量化模型，参数精度仍然是FP32，但是参数的数值范围是int8，导出的模型可以通过PaddleLite的opt模型转换工具完成模型转换。
+
+量化模型移动端部署的可参考 [移动端模型部署](../../lite/readme.md)
+
+备注：量化训练后的模型参数是float32类型，转inference model预测时相对不量化无加速效果，原因是量化后模型结构之间存在量化和反量化算子，如果要使用量化模型部署，建议使用TensorRT并设置precision为INT8加速量化模型的预测时间。

deploy/slim/quantization/README_en.md

@@ -0,0 +1,65 @@
+
+# PP-OCR Models Quantization
+
+Generally, a more complex model would achieve better performance in the task, but it also leads to some redundancy in the model.
+Quantization is a technique that reduces this redundancy by reducing the full precision data to a fixed number,
+so as to reduce model calculation complexity and improve model inference performance.
+
+This example uses PaddleSlim provided [APIs of Quantization](https://github.com/PaddlePaddle/PaddleSlim/blob/develop/docs/zh_cn/api_cn/dygraph/quanter/qat.rst) to compress the OCR model.
+
+It is recommended that you could understand following pages before reading this example：
+- [The training strategy of OCR model](../../../doc/doc_en/quickstart_en.md)
+- [PaddleSlim Document](https://github.com/PaddlePaddle/PaddleSlim/blob/develop/docs/zh_cn/api_cn/dygraph/quanter/qat.rst)
+
+## Quick Start
+Quantization is mostly suitable for the deployment of lightweight models on mobile terminals.
+After training, if you want to further compress the model size and accelerate the prediction, you can use quantization methods to compress the model according to the following steps.
+
+1. Install PaddleSlim
+2. Prepare trained model
+3. Quantization-Aware Training
+4. Export inference model
+5. Deploy quantization inference model
+
+
+### 1. Install PaddleSlim
+
+```bash
+pip3 install paddleslim==2.3.2
+```
+
+
+### 2. Download Pre-trained Model
+PaddleOCR provides a series of pre-trained [models](../../../doc/doc_en/models_list_en.md).
+If the model to be quantified is not in the list, you need to follow the [Regular Training](../../../doc/doc_en/quickstart_en.md) method to get the trained model.
+
+
+### 3. Quant-Aware Training
+Quantization training includes offline quantization training and online quantization training.
+Online quantization training is more effective. It is necessary to load the pre-trained model.
+After the quantization strategy is defined, the model can be quantified.
+
+The code for quantization training is located in `slim/quantization/quant.py`. For example, the training instructions of slim PPOCRv3 detection model are as follows:
+```
+# download provided model
+wget https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_distill_train.tar
+tar xf ch_PP-OCRv3_det_distill_train.tar
+
+python deploy/slim/quantization/quant.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o Global.pretrained_model='./ch_PP-OCRv3_det_distill_train/best_accuracy'   Global.save_model_dir=./output/quant_model_distill/
+```
+
+If you want to quantify the text recognition model, you can modify the configuration file and loaded model parameters.
+
+### 4. Export inference model
+
+Once we got the model after pruning and fine-tuning, we can export it as an inference model for the deployment of predictive tasks:
+
+```bash
+python deploy/slim/quantization/export_model.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o Global.checkpoints=output/quant_model/best_accuracy Global.save_inference_dir=./output/quant_inference_model
+```
+
+### 5. Deploy
+The numerical range of the quantized model parameters derived from the above steps is still FP32, but the numerical range of the parameters is int8.
+The derived model can be converted through the `opt tool` of PaddleLite.
+
+For quantitative model deployment, please refer to [Mobile terminal model deployment](../../lite/readme.md)

deploy/slim/quantization/export_model.py

@@ -0,0 +1,177 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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.
+
+import os
+import sys
+
+__dir__ = os.path.dirname(os.path.abspath(__file__))
+sys.path.append(__dir__)
+sys.path.insert(0, os.path.abspath(os.path.join(__dir__, '..', '..', '..')))
+sys.path.insert(
+    0, os.path.abspath(os.path.join(__dir__, '..', '..', '..', 'tools')))
+
+import argparse
+
+import paddle
+from paddle.jit import to_static
+
+from ppocr.modeling.architectures import build_model
+from ppocr.postprocess import build_post_process
+from ppocr.utils.save_load import load_model
+from ppocr.utils.logging import get_logger
+from tools.program import load_config, merge_config, ArgsParser
+from ppocr.metrics import build_metric
+import tools.program as program
+from paddleslim.dygraph.quant import QAT
+from ppocr.data import build_dataloader, set_signal_handlers
+from tools.export_model import export_single_model
+
+
+def main():
+    ############################################################################################################
+    # 1. quantization configs
+    ############################################################################################################
+    quant_config = {
+        # weight preprocess type, default is None and no preprocessing is performed. 
+        'weight_preprocess_type': None,
+        # activation preprocess type, default is None and no preprocessing is performed.
+        'activation_preprocess_type': None,
+        # weight quantize type, default is 'channel_wise_abs_max'
+        'weight_quantize_type': 'channel_wise_abs_max',
+        # activation quantize type, default is 'moving_average_abs_max'
+        'activation_quantize_type': 'moving_average_abs_max',
+        # weight quantize bit num, default is 8
+        'weight_bits': 8,
+        # activation quantize bit num, default is 8
+        'activation_bits': 8,
+        # data type after quantization, such as 'uint8', 'int8', etc. default is 'int8'
+        'dtype': 'int8',
+        # window size for 'range_abs_max' quantization. default is 10000
+        'window_size': 10000,
+        # The decay coefficient of moving average, default is 0.9
+        'moving_rate': 0.9,
+        # for dygraph quantization, layers of type in quantizable_layer_type will be quantized
+        'quantizable_layer_type': ['Conv2D', 'Linear'],
+    }
+    FLAGS = ArgsParser().parse_args()
+    config = load_config(FLAGS.config)
+    config = merge_config(config, FLAGS.opt)
+    logger = get_logger()
+    # build post process
+
+    post_process_class = build_post_process(config['PostProcess'],
+                                            config['Global'])
+
+    # build model
+    if hasattr(post_process_class, 'character'):
+        char_num = len(getattr(post_process_class, 'character'))
+        if config['Architecture']["algorithm"] in ["Distillation",
+                                                   ]:  # distillation model
+            for key in config['Architecture']["Models"]:
+                if config['Architecture']['Models'][key]['Head'][
+                        'name'] == 'MultiHead':  # for multi head
+                    if config['PostProcess'][
+                            'name'] == 'DistillationSARLabelDecode':
+                        char_num = char_num - 2
+                    # update SARLoss params
+                    assert list(config['Loss']['loss_config_list'][-1].keys())[
+                        0] == 'DistillationSARLoss'
+                    config['Loss']['loss_config_list'][-1][
+                        'DistillationSARLoss']['ignore_index'] = char_num + 1
+                    out_channels_list = {}
+                    out_channels_list['CTCLabelDecode'] = char_num
+                    out_channels_list['SARLabelDecode'] = char_num + 2
+                    config['Architecture']['Models'][key]['Head'][
+                        'out_channels_list'] = out_channels_list
+                else:
+                    config['Architecture']["Models"][key]["Head"][
+                        'out_channels'] = char_num
+        elif config['Architecture']['Head'][
+                'name'] == 'MultiHead':  # for multi head
+            if config['PostProcess']['name'] == 'SARLabelDecode':
+                char_num = char_num - 2
+            # update SARLoss params
+            assert list(config['Loss']['loss_config_list'][1].keys())[
+                0] == 'SARLoss'
+            if config['Loss']['loss_config_list'][1]['SARLoss'] is None:
+                config['Loss']['loss_config_list'][1]['SARLoss'] = {
+                    'ignore_index': char_num + 1
+                }
+            else:
+                config['Loss']['loss_config_list'][1]['SARLoss'][
+                    'ignore_index'] = char_num + 1
+            out_channels_list = {}
+            out_channels_list['CTCLabelDecode'] = char_num
+            out_channels_list['SARLabelDecode'] = char_num + 2
+            config['Architecture']['Head'][
+                'out_channels_list'] = out_channels_list
+        else:  # base rec model
+            config['Architecture']["Head"]['out_channels'] = char_num
+
+        if config['PostProcess']['name'] == 'SARLabelDecode':  # for SAR model
+            config['Loss']['ignore_index'] = char_num - 1
+
+    model = build_model(config['Architecture'])
+
+    # get QAT model
+    quanter = QAT(config=quant_config)
+    quanter.quantize(model)
+
+    load_model(config, model)
+
+    # build metric
+    eval_class = build_metric(config['Metric'])
+
+    # build dataloader
+    set_signal_handlers()
+    valid_dataloader = build_dataloader(config, 'Eval', device, logger)
+
+    use_srn = config['Architecture']['algorithm'] == "SRN"
+    model_type = config['Architecture'].get('model_type', None)
+    # start eval
+    metric = program.eval(model, valid_dataloader, post_process_class,
+                          eval_class, model_type, use_srn)
+    model.eval()
+
+    logger.info('metric eval ***************')
+    for k, v in metric.items():
+        logger.info('{}:{}'.format(k, v))
+
+    save_path = config["Global"]["save_inference_dir"]
+
+    arch_config = config["Architecture"]
+
+    if arch_config["algorithm"] == "SVTR" and arch_config["Head"][
+            "name"] != 'MultiHead':
+        input_shape = config["Eval"]["dataset"]["transforms"][-2][
+            'SVTRRecResizeImg']['image_shape']
+    else:
+        input_shape = None
+
+    if arch_config["algorithm"] in ["Distillation", ]:  # distillation model
+        archs = list(arch_config["Models"].values())
+        for idx, name in enumerate(model.model_name_list):
+            sub_model_save_path = os.path.join(save_path, name, "inference")
+            export_single_model(model.model_list[idx], archs[idx],
+                                sub_model_save_path, logger, input_shape,
+                                quanter)
+    else:
+        save_path = os.path.join(save_path, "inference")
+        export_single_model(model, arch_config, save_path, logger, input_shape,
+                            quanter)
+
+
+if __name__ == "__main__":
+    config, device, logger, vdl_writer = program.preprocess()
+    main()

deploy/slim/quantization/quant.py

@@ -0,0 +1,203 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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 os
+import sys
+
+__dir__ = os.path.dirname(os.path.abspath(__file__))
+sys.path.append(__dir__)
+sys.path.append(os.path.abspath(os.path.join(__dir__, '..', '..', '..')))
+sys.path.append(
+    os.path.abspath(os.path.join(__dir__, '..', '..', '..', 'tools')))
+
+import yaml
+import paddle
+import paddle.distributed as dist
+
+paddle.seed(2)
+
+from ppocr.data import build_dataloader, set_signal_handlers
+from ppocr.modeling.architectures import build_model
+from ppocr.losses import build_loss
+from ppocr.optimizer import build_optimizer
+from ppocr.postprocess import build_post_process
+from ppocr.metrics import build_metric
+from ppocr.utils.save_load import load_model
+import tools.program as program
+from paddleslim.dygraph.quant import QAT
+
+dist.get_world_size()
+
+
+class PACT(paddle.nn.Layer):
+    def __init__(self):
+        super(PACT, self).__init__()
+        alpha_attr = paddle.ParamAttr(
+            name=self.full_name() + ".pact",
+            initializer=paddle.nn.initializer.Constant(value=20),
+            learning_rate=1.0,
+            regularizer=paddle.regularizer.L2Decay(2e-5))
+
+        self.alpha = self.create_parameter(
+            shape=[1], attr=alpha_attr, dtype='float32')
+
+    def forward(self, x):
+        out_left = paddle.nn.functional.relu(x - self.alpha)
+        out_right = paddle.nn.functional.relu(-self.alpha - x)
+        x = x - out_left + out_right
+        return x
+
+
+quant_config = {
+    # weight preprocess type, default is None and no preprocessing is performed. 
+    'weight_preprocess_type': None,
+    # activation preprocess type, default is None and no preprocessing is performed.
+    'activation_preprocess_type': None,
+    # weight quantize type, default is 'channel_wise_abs_max'
+    'weight_quantize_type': 'channel_wise_abs_max',
+    # activation quantize type, default is 'moving_average_abs_max'
+    'activation_quantize_type': 'moving_average_abs_max',
+    # weight quantize bit num, default is 8
+    'weight_bits': 8,
+    # activation quantize bit num, default is 8
+    'activation_bits': 8,
+    # data type after quantization, such as 'uint8', 'int8', etc. default is 'int8'
+    'dtype': 'int8',
+    # window size for 'range_abs_max' quantization. default is 10000
+    'window_size': 10000,
+    # The decay coefficient of moving average, default is 0.9
+    'moving_rate': 0.9,
+    # for dygraph quantization, layers of type in quantizable_layer_type will be quantized
+    'quantizable_layer_type': ['Conv2D', 'Linear'],
+}
+
+
+def main(config, device, logger, vdl_writer):
+    # init dist environment
+    if config['Global']['distributed']:
+        dist.init_parallel_env()
+
+    global_config = config['Global']
+
+    # build dataloader
+    set_signal_handlers()
+    train_dataloader = build_dataloader(config, 'Train', device, logger)
+    if config['Eval']:
+        valid_dataloader = build_dataloader(config, 'Eval', device, logger)
+    else:
+        valid_dataloader = None
+
+    # build post process
+    post_process_class = build_post_process(config['PostProcess'],
+                                            global_config)
+
+    # build model
+    # for rec algorithm
+    if hasattr(post_process_class, 'character'):
+        char_num = len(getattr(post_process_class, 'character'))
+        if config['Architecture']["algorithm"] in ["Distillation",
+                                                   ]:  # distillation model
+            for key in config['Architecture']["Models"]:
+                if config['Architecture']['Models'][key]['Head'][
+                        'name'] == 'MultiHead':  # for multi head
+                    if config['PostProcess'][
+                            'name'] == 'DistillationSARLabelDecode':
+                        char_num = char_num - 2
+                    # update SARLoss params
+                    assert list(config['Loss']['loss_config_list'][-1].keys())[
+                        0] == 'DistillationSARLoss'
+                    config['Loss']['loss_config_list'][-1][
+                        'DistillationSARLoss']['ignore_index'] = char_num + 1
+                    out_channels_list = {}
+                    out_channels_list['CTCLabelDecode'] = char_num
+                    out_channels_list['SARLabelDecode'] = char_num + 2
+                    config['Architecture']['Models'][key]['Head'][
+                        'out_channels_list'] = out_channels_list
+                else:
+                    config['Architecture']["Models"][key]["Head"][
+                        'out_channels'] = char_num
+        elif config['Architecture']['Head'][
+                'name'] == 'MultiHead':  # for multi head
+            if config['PostProcess']['name'] == 'SARLabelDecode':
+                char_num = char_num - 2
+            # update SARLoss params
+            assert list(config['Loss']['loss_config_list'][1].keys())[
+                0] == 'SARLoss'
+            if config['Loss']['loss_config_list'][1]['SARLoss'] is None:
+                config['Loss']['loss_config_list'][1]['SARLoss'] = {
+                    'ignore_index': char_num + 1
+                }
+            else:
+                config['Loss']['loss_config_list'][1]['SARLoss'][
+                    'ignore_index'] = char_num + 1
+            out_channels_list = {}
+            out_channels_list['CTCLabelDecode'] = char_num
+            out_channels_list['SARLabelDecode'] = char_num + 2
+            config['Architecture']['Head'][
+                'out_channels_list'] = out_channels_list
+        else:  # base rec model
+            config['Architecture']["Head"]['out_channels'] = char_num
+
+        if config['PostProcess']['name'] == 'SARLabelDecode':  # for SAR model
+            config['Loss']['ignore_index'] = char_num - 1
+    model = build_model(config['Architecture'])
+
+    pre_best_model_dict = dict()
+    # load fp32 model to begin quantization
+    pre_best_model_dict = load_model(config, model, None, config['Architecture']["model_type"])
+
+    freeze_params = False
+    if config['Architecture']["algorithm"] in ["Distillation"]:
+        for key in config['Architecture']["Models"]:
+            freeze_params = freeze_params or config['Architecture']['Models'][
+                key].get('freeze_params', False)
+    act = None if freeze_params else PACT
+    quanter = QAT(config=quant_config, act_preprocess=act)
+    quanter.quantize(model)
+
+    if config['Global']['distributed']:
+        model = paddle.DataParallel(model)
+
+    # build loss
+    loss_class = build_loss(config['Loss'])
+
+    # build optim
+    optimizer, lr_scheduler = build_optimizer(
+        config['Optimizer'],
+        epochs=config['Global']['epoch_num'],
+        step_each_epoch=len(train_dataloader),
+        model=model)
+
+    # resume PACT training process
+    pre_best_model_dict = load_model(config, model, optimizer, config['Architecture']["model_type"])
+
+    # build metric
+    eval_class = build_metric(config['Metric'])
+
+    logger.info('train dataloader has {} iters, valid dataloader has {} iters'.
+                format(len(train_dataloader), len(valid_dataloader)))
+
+    # start train
+    program.train(config, train_dataloader, valid_dataloader, device, model,
+                  loss_class, optimizer, lr_scheduler, post_process_class,
+                  eval_class, pre_best_model_dict, logger, vdl_writer)
+
+
+if __name__ == '__main__':
+    config, device, logger, vdl_writer = program.preprocess(is_train=True)
+    main(config, device, logger, vdl_writer)

deploy/slim/quantization/quant_kl.py

@@ -0,0 +1,165 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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 os
+import sys
+
+__dir__ = os.path.dirname(os.path.abspath(__file__))
+sys.path.append(__dir__)
+sys.path.append(os.path.abspath(os.path.join(__dir__, '..', '..', '..')))
+sys.path.append(
+    os.path.abspath(os.path.join(__dir__, '..', '..', '..', 'tools')))
+
+import yaml
+import paddle
+import paddle.distributed as dist
+
+paddle.seed(2)
+
+from ppocr.data import build_dataloader, set_signal_handlers
+from ppocr.modeling.architectures import build_model
+from ppocr.losses import build_loss
+from ppocr.optimizer import build_optimizer
+from ppocr.postprocess import build_post_process
+from ppocr.metrics import build_metric
+from ppocr.utils.save_load import load_model
+import tools.program as program
+import paddleslim
+from paddleslim.dygraph.quant import QAT
+import numpy as np
+
+dist.get_world_size()
+
+
+class PACT(paddle.nn.Layer):
+    def __init__(self):
+        super(PACT, self).__init__()
+        alpha_attr = paddle.ParamAttr(
+            name=self.full_name() + ".pact",
+            initializer=paddle.nn.initializer.Constant(value=20),
+            learning_rate=1.0,
+            regularizer=paddle.regularizer.L2Decay(2e-5))
+
+        self.alpha = self.create_parameter(
+            shape=[1], attr=alpha_attr, dtype='float32')
+
+    def forward(self, x):
+        out_left = paddle.nn.functional.relu(x - self.alpha)
+        out_right = paddle.nn.functional.relu(-self.alpha - x)
+        x = x - out_left + out_right
+        return x
+
+
+quant_config = {
+    # weight preprocess type, default is None and no preprocessing is performed. 
+    'weight_preprocess_type': None,
+    # activation preprocess type, default is None and no preprocessing is performed.
+    'activation_preprocess_type': None,
+    # weight quantize type, default is 'channel_wise_abs_max'
+    'weight_quantize_type': 'channel_wise_abs_max',
+    # activation quantize type, default is 'moving_average_abs_max'
+    'activation_quantize_type': 'moving_average_abs_max',
+    # weight quantize bit num, default is 8
+    'weight_bits': 8,
+    # activation quantize bit num, default is 8
+    'activation_bits': 8,
+    # data type after quantization, such as 'uint8', 'int8', etc. default is 'int8'
+    'dtype': 'int8',
+    # window size for 'range_abs_max' quantization. default is 10000
+    'window_size': 10000,
+    # The decay coefficient of moving average, default is 0.9
+    'moving_rate': 0.9,
+    # for dygraph quantization, layers of type in quantizable_layer_type will be quantized
+    'quantizable_layer_type': ['Conv2D', 'Linear'],
+}
+
+
+def sample_generator(loader):
+    def __reader__():
+        for indx, data in enumerate(loader):
+            images = np.array(data[0])
+            yield images
+
+    return __reader__
+
+def sample_generator_layoutxlm_ser(loader):
+    def __reader__():
+        for indx, data in enumerate(loader):
+            input_ids = np.array(data[0])
+            bbox = np.array(data[1])
+            attention_mask = np.array(data[2])
+            token_type_ids = np.array(data[3])
+            images = np.array(data[4])
+            yield [input_ids, bbox, attention_mask, token_type_ids, images]
+
+    return __reader__
+
+def main(config, device, logger, vdl_writer):
+    # init dist environment
+    if config['Global']['distributed']:
+        dist.init_parallel_env()
+
+    global_config = config['Global']
+
+    # build dataloader
+    set_signal_handlers()
+    config['Train']['loader']['num_workers'] = 0
+    is_layoutxlm_ser =  config['Architecture']['model_type'] =='kie' and config['Architecture']['Backbone']['name'] == 'LayoutXLMForSer'
+    train_dataloader = build_dataloader(config, 'Train', device, logger)
+    if config['Eval']:
+        config['Eval']['loader']['num_workers'] = 0
+        valid_dataloader = build_dataloader(config, 'Eval', device, logger)
+        if is_layoutxlm_ser:
+            train_dataloader = valid_dataloader
+    else:
+        valid_dataloader = None
+
+    paddle.enable_static()
+    exe = paddle.static.Executor(device)
+
+    if 'inference_model' in global_config.keys():  # , 'inference_model'):
+        inference_model_dir = global_config['inference_model']
+    else:
+        inference_model_dir = os.path.dirname(global_config['pretrained_model'])
+        if  not (os.path.exists(os.path.join(inference_model_dir, "inference.pdmodel")) and \
+            os.path.exists(os.path.join(inference_model_dir, "inference.pdiparams")) ):
+            raise ValueError(
+                "Please set inference model dir in Global.inference_model or Global.pretrained_model for post-quantazition"
+            )
+    
+    if is_layoutxlm_ser:
+        generator = sample_generator_layoutxlm_ser(train_dataloader)
+    else:
+        generator = sample_generator(train_dataloader)
+
+    paddleslim.quant.quant_post_static(
+        executor=exe,
+        model_dir=inference_model_dir,
+        model_filename='inference.pdmodel',
+        params_filename='inference.pdiparams',
+        quantize_model_path=global_config['save_inference_dir'],
+        sample_generator=generator,
+        save_model_filename='inference.pdmodel',
+        save_params_filename='inference.pdiparams',
+        batch_size=1,
+        batch_nums=None)
+
+
+if __name__ == '__main__':
+    config, device, logger, vdl_writer = program.preprocess(is_train=True)
+    main(config, device, logger, vdl_writer)