I have tried the bytetrack_x_mot17.pth.tar/756MB model in detecting persons, i found that this tar-file has a much better performance than the original yolox_x.pth/756MB from Megvii-BaseDetection.
Would you please to show some details about the training of these Standard models?
such as what type of GPU? how many GPUs? how long was the training? were there any freezed layers during the finetune of yolox_x.pth? .....
Thank you for this amazing project!
I have tried the bytetrack_x_mot17.pth.tar/756MB model in detecting persons, i found that this tar-file has a much better performance than the original yolox_x.pth/756MB from Megvii-BaseDetection.
Would you please to show some details about the training of these Standard models? such as what type of GPU? how many GPUs? how long was the training? were there any freezed layers during the finetune of yolox_x.pth? .....