search

yjh0410 / SAMI

Masked AutoEncoders leveraging Segment-Anything
12 stars 2 forks source link

Confused #3

Open yyy1998-i opened 2 months ago

yyy1998-i commented 2 months ago

Hello author, I used different teacher models in pre-training, namely medsam_checkpoint,sam_vit_b_checkpoint, and the other one did not use the teacher model. As follows, I found that different teacher models differ greatly in pre-training. medsam_checkpoint loss is 18.9 at the beginning and 4.2 after 300 epoches training, which is not as good as not using the teacher model at the beginning. Because my data is medical images, the structure of sam_vit_b_checkpoint model and medsam_checkpoint model are exactly the same, which makes me feel confused. I hope you can give me some tips, thank you for your reply fc4a00a5-8789-4885-a133-cbf519262ed2 992fc17c843f1218ce3cd3044fd670e4 386bf12575cc518fbd8cb5ad7b42c7e7

yyy1998-i commented 2 months ago

https://drive.google.com/drive/folders/1ETWmi4AiniJeWOt6HAsYgTjYv_fkgzoN?usp=drive_link med_sam_vit_b checkpoint

lpc-97 commented 5 days ago

你好 请问你后面 训练的效果如何呢 微调之后 测试结果有多少呢