Tao model benchmarking with model from other repository

Please provide the following information when requesting support.

• Hardware : A10
• Network Type : Yolo_v4_csp_darknet53

I have trained Yolo_v4_csp_darknet53 from GitHub - WongKinYiu/ScaledYOLOv4: Scaled-YOLOv4: Scaling Cross Stage Partial Network and using Tao toolkit on same dataset (custom data). In Tao toolkit and above repository I followed default training configurations but weirdly I am getting two different results on both models. Tao model suffers badly missing common scenarios which is easily identified by model trained on native repository.
I trained both models for nearly 200 epochs. I am not getting why this is happening. Please help !!!

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

May I know more info about the results? Thanks.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.