DAMO-YOLO: a fast and accurate object detection method with some new techs, including NAS backbones, efficient RepGFPN, ZeroHead, AlignedOTA, and distillation enhancement.
[X] I have read the README carefully. 我已经仔细阅读了README上的操作指引。
[X] I want to train my custom dataset, and I have read the tutorials for finetune on your data carefully and organize my dataset correctly; 我想训练自定义数据集，我已经仔细阅读了训练自定义数据的教程，以及按照正确的目录结构存放数据集。
[X] I have pulled the latest code of main branch to run again and the problem still existed. 我已经拉取了主分支上最新的代码，重新运行之后，问题仍不能解决。
Search before asking
[X] I have searched the DAMO-YOLO issues and found no similar questions.
Question
Is it okay to take weight from augmentation epochs?
Say, during training my model performs best during augmentation epochs.
But when MOSAIC is off, and trained without augmentations, the score does not fall rather keeps decreasing.
To deploy in real life, should I choose the weights from MOSAIC augmentation epochs even though it is the best?
Before Asking
[X] I have read the README carefully. 我已经仔细阅读了README上的操作指引。
[X] I want to train my custom dataset, and I have read the tutorials for finetune on your data carefully and organize my dataset correctly; 我想训练自定义数据集，我已经仔细阅读了训练自定义数据的教程，以及按照正确的目录结构存放数据集。
[X] I have pulled the latest code of main branch to run again and the problem still existed. 我已经拉取了主分支上最新的代码，重新运行之后，问题仍不能解决。
Search before asking
Question
Is it okay to take weight from augmentation epochs? Say, during training my model performs best during augmentation epochs. But when MOSAIC is off, and trained without augmentations, the score does not fall rather keeps decreasing.
To deploy in real life, should I choose the weights from MOSAIC augmentation epochs even though it is the best?
Additional
No response