RTMPose TensorRT Deployment - C++ Pre and Post-processing Code Example RTMPose TensorRT 部署 - C++ 前后处理代码样例

[OpenMMLab-Assistant-004]

RTMPose TensorRT Deployment - C++ Pre and Post-processing Code Example

RTMPose TensorRT 部署 - C++ 前后处理代码样例

TensorRT Deployment Framework for RTMPose - C++ Pre and Post-processing Code Example
To demonstrate C++ inference of the RTMPose model using the TensorRT deployment framework, follow these steps in the projects/rtmpose/examples/{deployment_framework} directory (create the directory if it doesn't exist):

1. Read Input Image
2. Define a preprocess() function to showcase the pre-processing steps, including computing center, scale, and image resizing. Refer to the val_pipeline in the RTMPose config for image pre-processing (core code in GetBBoxCenterScale and TopdownAffine).
3. Define a postprocess() function to showcase the post-processing steps, inspired by SimCCLabel's decode() function in mmpose/codec/simcc_label.py.
4. Provide a method for model export. This example assumes loading the model from a local path.
5. Maintain code readability by including sufficient comments to facilitate understanding and learning, especially for beginners.
6. Use a README file to provide additional explanations and instructions.

基于 TensorRT 部署框架，在 projects/rtmpose/examples/{部署框架名}/目录（没有则新建）中演示 C++ RTMPose 模型推理，包含以下内容：

1. 读取输入图片
2. 定义一个 preprocess() 展示预处理过程（计算center, scale，图片缩放），参考 RTMPose config 中的 val_pipeline 进行图片预处理（核心代码在GetBBoxCenterScale和TopdownAffine ）
3. 定义一个 postprocess() 展示后处理过程，参考 mmpose/codec/simcc_label.py中，SimCCLabel的decode()
4. 请给出模型导出的方式，该样例从本地路径读取模型
5. 尽量保证代码的可读性，尽可能多地进行注释，方便新手理解和学习
6. 可以使用 README 进行说明

Task Information

Repository：MMPose

Difficulty：Medium

Points：30

Estimated Time Consumption：14 days

Task Status

Claimant's Github ID：Dominic23331

Claim Time：2023年05月15日

PR Link：open-mmlab/mmpose#2397

Points to Note

To claim the task, please reply Claim the mission in the comment section. After claiming the task, please add our assistant bot called MeowMewo on WeChat at: openmmlabwx.

After completing the task, please submit a PR with the title starting with [MMSIG]. After submission, please reply with your PR link in the comments section of this Discussion. We will review your work as soon as possible.

在评论区回复 Claim the mission 领取任务，领取任务后请及时添加小助手喵喵微信: openmmlabwx

完成任务后提交相关链接（如是PR，标题请以[MMSIG]开头）回复到本Disucssion评论区，我们会第一时间审核您的贡献。