Wang Ao

Thanks. Could you please manually set `self.is_coco = True` for the `validator` to keep the same way as COCO for calculating mAP?

Thanks for the detailed explanation. Would you mind sharing your validation set and the checkpoint with us? We are trying to investigate this and confirm if there is an issue...

Thank you! Our email is [email protected].

Thanks for your interest! It seems that there is no `torch.topk` in yolov8. Could you please confirm that? Thank you!

Thanks. This issue may rely on ncnn to add the support for `torch.topk`.

Thank you for your great efforts! Due to that YOLOv10 does not rely on the NMS, it can be directly exported to end2end onnx/TensorRT without the NMS plugin, like https://github.com/THU-MIG/yolov10?tab=readme-ov-file#export

Thanks for your interest! Please feel free to reopen this issue if you have further questions.

Thanks for your interest! Could you please provide more details? Do you use the updated [checkpoints](https://github.com/THU-MIG/yolov10/releases/tag/v1.1)? The updated checkpoints include the training hyperparameters. May that have an impact?

@Godk02 @Sunny20236 Thanks. 1. Are the training hyper-parameters, e.g., data augmentation hyperparameters, the same for yolov10x and yolov8x? 2. Could you please check the performance gap between the one-to-one head...

@salwaghanim Thanks for your detailed explanation and great help! Could you please provide more details about that image resizing is different while training and inference in yolov10? We thought that...