I have implemented Yolo on custom dataset...and want to deploy it...exported model (best.torchscript.ptl) is working in demo app without getting any error...However, it provides very inaccurate prediction...so, I tried to upload original model (best.pt)...and it's not working...can we not directly work with original pytorch model generated after training completing training process..?
Next issue is exported model has large size compared to trained model...
I look forward to Your response...
I have implemented Yolo on custom dataset...and want to deploy it...exported model (best.torchscript.ptl) is working in demo app without getting any error...However, it provides very inaccurate prediction...so, I tried to upload original model (best.pt)...and it's not working...can we not directly work with original pytorch model generated after training completing training process..? Next issue is exported model has large size compared to trained model... I look forward to Your response...
Thank You...