I am using five types(like, fist,palm, peace and no_gesture) to train my own SSDLiteMobileNetV3Large model. My config file is here
myconfig.txt
And here is how I place my data and annotations.
After training for two whole days, my model only achieves a mAP of around 0.4-0.5. Additionally, when I use the trained .pth file with demo.py, it can only detect the fist gesture.
Can someone help me identify what might be going wrong? Thanks!
I am using five types(like, fist,palm, peace and no_gesture) to train my own SSDLiteMobileNetV3Large model. My config file is here myconfig.txt
And here is how I place my data and annotations.
After training for two whole days, my model only achieves a mAP of around 0.4-0.5. Additionally, when I use the trained .pth file with demo.py, it can only detect the fist gesture.
Can someone help me identify what might be going wrong? Thanks!