fan23j/yolov5-vitpose-video-annotator

support onnx/tensorrt

Closed this issue · 2 comments

wanghlsara commented

Hi,
I have another question for you.
Your ViTPose with Halpe model works well with Pytorch. I sucessfully exported some official vitpose models to onnx/tensorrt file by mmdeploy(command is like this), but I don't know how to export from your Haple model. Could you share me some more infomation about your Halpe model?

fan23j commented

since i developed from the ViTPose repo, the export script should work. If you provide the config file with the weights, I believe it should work. Unless you already ran into an error?

wanghlsara commented

Thanks a lot for your reply.
I finally exported your halpe .pth model to tensorrt .engine file and it can infer very well now.
I will close this issue, and maybe ask for your help with training techniques in the future.