Megvii-BaseDetection / Megvii-BaseDetection/YOLOX
torch2trt 转换问题
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 10.6k
- Forks
- 2.5k
- PR merge metrics
- No merged PRs in 30d
Description
我在使用torch2trt转我的yolox模型的时候，原本可以转，但是我是用repvggblock替换原本的cspblock的时候，torch2trt就会失效，不会出现任何错误而是直接中止转换。就像这样(yolox) R:\YOLOX>python tool/trt.py --onlypth
2023-12-07 20:53:47.909 | INFO | main:main:63 - loaded checkpoint done.
[12/07/2023-20:53:50] [TRT] [I] [MemUsageChange] Init CUDA: CPU +776, GPU +0, now: CPU 13024, GPU 2274 (MiB)
(yolox) R:\YOLOX>。但是使用pth to onnx 再从onnx to engine是可行的，但是不会生成model_trt文件。能否告知一下为什么torch2trt不支持repvggblock。或者能提供一个调用python api的推理脚本吗，方便使用engine文件。十分感谢大佬！
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading and running tool/trt.py with the reported YOLOX checkpoint and the RepVGGBlock replacement, then compare the torch2trt path with the working ONNX-to-engine path. Done means identifying why conversion stops without an error and either documenting a supported conversion path or providing the requested Python API inference script for the engine file.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- devtools, machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100