Open shermannws opened 3 years ago
I have trained the model successfully following these instructions
https://github.com/NVIDIA-AI-IOT/trt_pose/issues/59#issuecomment-668252922
If you can annotate your data in COCO format it should not be a problem
I am using TRT_Pose and realised that the current model could not detect if the human is lying down. I am planning to train a new model. Have anyone done custom training before? What should I do and what data would I have to prepare? Any help would be appreciated! Thank you so much!