107 lines
7.6 KiB
Markdown
107 lines
7.6 KiB
Markdown
|
# Linux GPU/CPU PACT量化训练推理测试
|
|||
|
|
|||
|
Linux GPU/CPU PACT量化训练推理测试的主程序为`test_train_inference_python.sh`,可以测试基于Python的模型PACT在线量化等基本功能。
|
|||
|
|
|||
|
## 1. 测试结论汇总
|
|||
|
|
|||
|
- 训练相关:
|
|||
|
|
|||
|
| 算法名称 | 模型名称 | 单机单卡 |
|
|||
|
| :-------------: | :-------------------------------------: | :----------: |
|
|||
|
| MobileNetV3 | MobileNetV3_large_x1_0 | PACT量化训练 |
|
|||
|
| PP-ShiTu | GeneralRecognition_PPLCNet_x2_5 | PACT量化训练 |
|
|||
|
| PPHGNet | PPHGNet_small | PACT量化训练 |
|
|||
|
| PPHGNet | PPHGNet_tiny | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x0_25 | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x0_35 | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x0_5 | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x0_75 | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x1_0 | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x1_5 | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x2_0 | PACT量化训练 |
|
|||
|
| PPLCNet | PPLCNet_x2_5 | PACT量化训练 |
|
|||
|
| PPLCNetV2 | PPLCNetV2_base | PACT量化训练 |
|
|||
|
| ResNet | ResNet50 | PACT量化训练 |
|
|||
|
| ResNet | ResNet50_vd | PACT量化训练 |
|
|||
|
| SwinTransformer | SwinTransformer_tiny_patch4_window7_224 | PACT量化训练 |
|
|||
|
|
|||
|
- 推理相关:
|
|||
|
|
|||
|
| 算法名称 | 模型名称 | device_CPU | device_GPU | batchsize |
|
|||
|
| :-------------: | :-------------------------------------: | :--------: | :--------: | :-------: |
|
|||
|
| MobileNetV3 | MobileNetV3_large_x1_0 | 支持 | 支持 | 1 |
|
|||
|
| PP-ShiTu | GeneralRecognition_PPLCNet_x2_5 | 支持 | 支持 | 1 |
|
|||
|
| PPHGNet | PPHGNet_small | 支持 | 支持 | 1 |
|
|||
|
| PPHGNet | PPHGNet_tiny | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x0_25 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x0_35 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x0_5 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x0_75 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x1_0 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x1_5 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x2_0 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNet | PPLCNet_x2_5 | 支持 | 支持 | 1 |
|
|||
|
| PPLCNetV2 | PPLCNetV2_base | 支持 | 支持 | 1 |
|
|||
|
| ResNet | ResNet50 | 支持 | 支持 | 1 |
|
|||
|
| ResNet | ResNet50_vd | 支持 | 支持 | 1 |
|
|||
|
| SwinTransformer | SwinTransformer_tiny_patch4_window7_224 | 支持 | 支持 | 1 |
|
|||
|
|
|||
|
## 2. 测试流程
|
|||
|
|
|||
|
一下测试流程以 MobileNetV3_large_x1_0 模型为例。
|
|||
|
|
|||
|
### 2.1 准备环境
|
|||
|
|
|||
|
- 安装PaddlePaddle:如果您已经安装了2.2或者以上版本的paddlepaddle,那么无需运行下面的命令安装paddlepaddle。
|
|||
|
```bash
|
|||
|
# 需要安装2.2及以上版本的Paddle
|
|||
|
# 安装GPU版本的Paddle
|
|||
|
python3.7 -m pip install paddlepaddle-gpu==2.2.0
|
|||
|
# 安装CPU版本的Paddle
|
|||
|
python3.7 -m pip install paddlepaddle==2.2.0
|
|||
|
```
|
|||
|
|
|||
|
- 安装PaddleSlim
|
|||
|
```bash
|
|||
|
python3.7 -m pip install paddleslim==2.2.0
|
|||
|
```
|
|||
|
|
|||
|
- 安装依赖
|
|||
|
```bash
|
|||
|
python3.7 -m pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
|
|||
|
```
|
|||
|
|
|||
|
- 安装AutoLog(规范化日志输出工具)
|
|||
|
```bash
|
|||
|
python3.7 -m pip install https://paddleocr.bj.bcebos.com/libs/auto_log-1.2.0-py3-none-any.whl
|
|||
|
```
|
|||
|
|
|||
|
### 2.2 准备数据和模型
|
|||
|
|
|||
|
```bash
|
|||
|
bash test_tipc/prepare.sh test_tipc/config/PPLCNetV2/PPLCNetV2_base_train_pact_infer_python.txt lite_train_lite_infer
|
|||
|
```
|
|||
|
|
|||
|
在线量化的操作流程,可参考[文档](../../deploy/slim/README.md)。
|
|||
|
|
|||
|
### 2.3 功能测试
|
|||
|
|
|||
|
以`MobileNetV3_large_x1_0`的`Linux GPU/CPU PACT在线量化训练推理测试`为例,命令如下所示。
|
|||
|
|
|||
|
```bash
|
|||
|
bash test_tipc/test_train_inference_python.sh test_tipc/config/PPLCNetV2/PPLCNetV2_base_train_pact_infer_python.txt lite_train_lite_infer
|
|||
|
```
|
|||
|
|
|||
|
输出结果如下,表示命令运行成功。
|
|||
|
|
|||
|
```log
|
|||
|
Run successfully with command - MobileNetV3_large_x1_0 - python3.7 tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False -o Slim.quant.name=pact -o Optimizer.lr.learning_rate=0.01 -o Global.device=gpu -o Global.output_dir=./test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/pact_train_gpus_0_autocast_null -o Global.epochs=2 -o DataLoader.Train.sampler.batch_size=8 !
|
|||
|
Run successfully with command - MobileNetV3_large_x1_0 - python3.7 tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml -o Slim.quant.name=pact -o Global.pretrained_model=./test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/pact_train_gpus_0_autocast_null/MobileNetV3_large_x1_0/latest -o Global.device=gpu !
|
|||
|
Run successfully with command - MobileNetV3_large_x1_0 - python3.7 tools/export_model.py -c ppcls/configs/slim/MobileNetV3_large_x1_0_quantization.yaml -o Slim.quant.name=pact -o Global.pretrained_model=./test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/pact_train_gpus_0_autocast_null/MobileNetV3_large_x1_0/latest -o Global.save_inference_dir=./test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/pact_train_gpus_0_autocast_null!
|
|||
|
Run successfully with command - MobileNetV3_large_x1_0 - python3.7 python/predict_cls.py -c configs/inference_cls.yaml -o Global.use_gpu=True -o Global.use_tensorrt=False -o Global.use_fp16=False -o Global.inference_model_dir=.././test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/pact_train_gpus_0_autocast_null -o Global.batch_size=1 -o Global.infer_imgs=../dataset/ILSVRC2012/val -o Global.benchmark=True > .././test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/infer_gpu_usetrt_False_precision_False_batchsize_1.log 2>&1 !
|
|||
|
Run successfully with command - MobileNetV3_large_x1_0 - python3.7 python/predict_cls.py -c configs/inference_cls.yaml -o Global.use_gpu=False -o Global.enable_mkldnn=False -o Global.cpu_num_threads=1 -o Global.inference_model_dir=.././test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/pact_train_gpus_0_autocast_null -o Global.batch_size=1 -o Global.infer_imgs=../dataset/ILSVRC2012/val -o Global.benchmark=True > .././test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer/infer_cpu_usemkldnn_False_threads_1_batchsize_1.log 2>&1 !
|
|||
|
```
|
|||
|
同时,测试过程中的日志保存在`PaddleClas/test_tipc/output/MobileNetV3_large_x1_0/lite_train_lite_infer`下。
|
|||
|
|
|||
|
如果运行失败,也会在终端中输出运行失败的日志信息以及对应的运行命令。可以基于该命令,分析运行失败的原因。
|