Jiarui Yu
Jiarui Yu
Hello! I have some questions about settings. I notice that the default training script does not use **normalize_prefix**: `python train.py --only_prefix --data ./data/coco/oscar_split_ViT-B_32_train.pkl --out_dir ./coco_train/ --mapping_type transformer --num_layres 8 --prefix_length...
Hi @mt-cly, please refer to the pull request #135, it fixes few bugs and may help.
@wade0604 @uobinxiao @rocketche I fix some bugs, set random seed to ensure reproducibility and evaluate the new code on ISIC dataset. Please refer to PR #135.
@sijie-Xu #135 may help, please check~
Hi @WuJunde, junde please merge these commits into the main branch. And also suggested to check similar bugs in your recent repo [medical sam2](https://github.com/SuperMedIntel/Medical-SAM2.git). Feel free to contact!
+1,looking forward to the author’s improvement @hiyouga
@panboshui This may caused by some bugs, please refer to Pull Request #135 and it may help.