[Bug] YOLOX transform fp16 onnx

open-mmlab / mmdeploy

OpenMMLab Model Deployment Framework

https://mmdeploy.readthedocs.io/en/latest/

Apache License 2.0

2.77k stars 636 forks source link

[Bug] YOLOX transform fp16 onnx #2822

Open CH-Echo opened 2 months ago

CH-Echo commented 2 months ago

Checklist

[x] I have searched related issues but cannot get the expected help.
[X] 2. I have read the FAQ documentation but cannot get the expected help.
[X] 3. The bug has not been fixed in the latest version.

Describe the bug

I would like to ask if Yolox in RTMpose does not convert the configuration file corresponding to fp16. How should I convert it?

Reproduction

It prompts that the transfer to onnx fails, which BatchNMS module does not support it?

Environment

normal

Error traceback

No response

CH-Echo commented 2 months ago

[W shape_type_inference.cpp:1972] Warning: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function. (function UpdateReliable) [W shape_type_inference.cpp:1972] Warning: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function. (function UpdateReliable) 09/04 18:06:56 - mmengine - INFO - Execute onnx optimize passes. 09/04 18:06:56 - mmengine - WARNING - Can not optimize model, please build torchscipt extension.