mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2025-10-31 20:02:53 +08:00
104d965b38
* add yolov5 and ppyoloe for rk1126 * update code, rename rk1126 to rv1126 * add PP-Liteseg * update lite lib * updade doc for PPYOLOE * update doc * fix docs * fix doc and examples * update code * uodate doc * update doc Co-authored-by: Jason <jiangjiajun@baidu.com>
51 lines
5.9 KiB
Markdown
Executable File
51 lines
5.9 KiB
Markdown
Executable File
# PaddleClas 量化模型部署
|
|
FastDeploy已支持部署量化模型,并提供一键模型自动化压缩的工具.
|
|
用户可以使用一键模型自动化压缩工具,自行对模型量化后部署, 也可以直接下载FastDeploy提供的量化模型进行部署.
|
|
|
|
## FastDeploy一键模型自动化压缩工具
|
|
FastDeploy 提供了一键模型自动化压缩工具, 能够简单地通过输入一个配置文件, 对模型进行量化.
|
|
详细教程请见: [一键模型自动化压缩工具](../../../../../tools/common_tools/auto_compression/)
|
|
注意: 推理量化后的分类模型仍然需要FP32模型文件夹下的inference_cls.yaml文件, 自行量化的模型文件夹内不包含此yaml文件, 用户从FP32模型文件夹下复制此yaml文件到量化后的模型文件夹内即可。
|
|
|
|
## 下载量化完成的PaddleClas模型
|
|
用户也可以直接下载下表中的量化模型进行部署.
|
|
|
|
Benchmark表格说明:
|
|
- Runtime时延为模型在各种Runtime上的推理时延,包含CPU->GPU数据拷贝,GPU推理,GPU->CPU数据拷贝时间. 不包含模型各自的前后处理时间.
|
|
- 端到端时延为模型在实际推理场景中的时延, 包含模型的前后处理.
|
|
- 所测时延均为推理1000次后求得的平均值, 单位是毫秒.
|
|
- INT8 + FP16 为在推理INT8量化模型的同时, 给Runtime 开启FP16推理选项
|
|
- INT8 + FP16 + PM, 为在推理INT8量化模型和开启FP16的同时, 开启使用Pinned Memory的选项,可加速GPU->CPU数据拷贝的速度
|
|
- 最大加速比, 为FP32时延除以INT8推理的最快时延,得到最大加速比.
|
|
- 策略为量化蒸馏训练时, 采用少量无标签数据集训练得到量化模型, 并在全量验证集上验证精度, INT8精度并不代表最高的INT8精度.
|
|
- CPU为Intel(R) Xeon(R) Gold 6271C, 所有测试中固定CPU线程数为1. GPU为Tesla T4, TensorRT版本8.4.15.
|
|
|
|
### Runtime Benchmark
|
|
| 模型 |推理后端 |部署硬件 | FP32 Runtime时延 | INT8 Runtime时延 | INT8 + FP16 Runtime时延 | INT8+FP16+PM Runtime时延 | 最大加速比 | FP32 Top1 | INT8 Top1 | 量化方式 |
|
|
| ------------------- | -----------------|-----------| -------- |-------- |-------- | --------- |-------- |----- |----- |----- |
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | TensorRT | GPU | 3.55 | 0.99|0.98|1.06 | 3.62 | 79.12 | 79.06 | 离线量化 |
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | Paddle-TensorRT | GPU | 3.46 |None |0.87|1.03 | 3.98 | 79.12 | 79.06 | 离线量化 |
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | ONNX Runtime | CPU | 76.14 | 35.43 |None|None | 2.15 | 79.12 | 78.87| 离线量化|
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | Paddle Inference | CPU | 76.21 | 24.01 |None|None | 3.17 | 79.12 | 78.55 | 离线量化|
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | TensorRT | GPU | 0.91 | 0.43 |0.49 | 0.54 | 2.12 |77.89 | 76.86 | 离线量化 |
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | Paddle-TensorRT | GPU | 0.88| None| 0.49|0.51 | 1.80 |77.89 | 76.86 | 离线量化 |
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | ONNX Runtime | CPU | 30.53 | 9.59|None|None | 3.18 |77.89 | 75.09 |离线量化 |
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | Paddle Inference | CPU | 12.29 | 4.68 | None|None|2.62 |77.89 | 71.36 |离线量化 |
|
|
|
|
### 端到端 Benchmark
|
|
| 模型 |推理后端 |部署硬件 | FP32 End2End时延 | INT8 End2End时延 | INT8 + FP16 End2End时延 | INT8+FP16+PM End2End时延 | 最大加速比 | FP32 Top1 | INT8 Top1 | 量化方式 |
|
|
| ------------------- | -----------------|-----------| -------- |-------- |-------- | --------- |-------- |----- |----- |----- |
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | TensorRT | GPU | 4.92| 2.28|2.24|2.23 | 2.21 | 79.12 | 79.06 | 离线量化 |
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | Paddle-TensorRT | GPU | 4.48|None |2.09|2.10 | 2.14 | 79.12 | 79.06 | 离线量化 |
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | ONNX Runtime | CPU | 77.43 | 41.90 |None|None | 1.85 | 79.12 | 78.87| 离线量化|
|
|
| [ResNet50_vd](https://bj.bcebos.com/paddlehub/fastdeploy/resnet50_vd_ptq.tar) | Paddle Inference | CPU | 80.60 | 27.75 |None|None | 2.90 | 79.12 | 78.55 | 离线量化|
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | TensorRT | GPU | 2.19 | 1.48|1.57| 1.57 | 1.48 |77.89 | 76.86 | 离线量化 |
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | Paddle-TensorRT | GPU | 2.04| None| 1.47|1.45 | 1.41 |77.89 | 76.86 | 离线量化 |
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | ONNX Runtime | CPU | 34.02 | 12.97|None|None | 2.62 |77.89 | 75.09 |离线量化 |
|
|
| [MobileNetV1_ssld](https://bj.bcebos.com/paddlehub/fastdeploy/mobilenetv1_ssld_ptq.tar) | Paddle Inference | CPU | 16.31 | 7.42 | None|None| 2.20 |77.89 | 71.36 |离线量化 |
|
|
|
|
## 详细部署文档
|
|
|
|
- [Python部署](python)
|
|
- [C++部署](cpp)
|