The mapping is a big issue in the model prediction

[chenscottus]

Hi Jack,

 We have the similliar results like your examples: 
    https://github.com/fan23j/yolov7-pose-whole-body/tree/main/onnx_inference/

The mapping is a big issue in the model prediction.

 Here are more examples from our training ( Left side is ground true, right side is prediction):

     ![pose_wb_comparison](https://user-images.githubusercontent.com/5216043/234075684-7b2fbd9b-422a-4d4f-afad-56ac8ac4e5a2.jpg)

[chenscottus]

https://user-images.githubusercontent.com/5216043/234075684-7b2fbd9b-422a-4d4f-afad-56ac8ac4e5a2.jpg

[fan23j]

Am training another model to see if I can debug the issue

[chenscottus]

Here is the result after we re-training the model:

There is Only a little improvement in performance but the results are very poor.

Thanks,

-Scott

[fan23j]

What is the mAP after training? Do the images drawn in mosaic through test.py also present this mapping issue? I presume it's not a performance issue but rather a mapping one.

[chenscottus]

it was somewhere around 0.78 after training. The images drawn in mosaic through test.py also present this mapping issue.

On Mon, May 1, 2023 at 6:20 PM Jack Fan @.***> wrote:

What is the mAP after training? Do the images drawn in mosaic through test.py also present this mapping issue? I presume it's not a performance issue but rather a mapping one.

— Reply to this email directly, view it on GitHub https://github.com/fan23j/yolov7-pose-whole-body/issues/4#issuecomment-1530743199, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABHZOK3ECLZ4NTCIP72A52DXEBOPVANCNFSM6AAAAAAXJ5DT4M . You are receiving this because you authored the thread.Message ID: @.***>

[fan23j]

5