search

facebookresearch / segment-anything-2

The repository provides code for running inference with the Meta Segment Anything Model 2 (SAM 2), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.
Apache License 2.0
10.64k stars 849 forks source link

Can you post the training code? #115

Open LiGuo12 opened 1 month ago

LiGuo12 commented 1 month ago

The forward function in the SAM2Base class is currently empty, and the loss function used is not specified. Could you please release the complete training code? This would allow the community to verify your results and perform fine-tuning. Your contributions and assistance are greatly appreciated.

bhack commented 1 month ago

It was already requested here for the prompts simulator https://github.com/facebookresearch/segment-anything-2/issues/114

and more in general here for the training/fine-tuning code https://github.com/facebookresearch/segment-anything-2/issues/12

JunMa11 commented 1 month ago

Hi @LiGuo12 @bhack Here is the fine-tuning code for image segmentation https://github.com/bowang-lab/MedSAM/blob/MedSAM2/README.md

LiGuo12 commented 1 month ago

Hi @LiGuo12 @bhack Here is the fine-tuning code for image segmentation https://github.com/bowang-lab/MedSAM/blob/MedSAM2/README.md

Hi @JunMa11, Thanks for the wonderful code. Looking forward to your fine-tuning weights on the medical dataset.

bhack commented 1 month ago

The most interesting and missing part is to fine-tune the decoder/prompt encoder on the video mode

EricLina commented 1 month ago

The most interesting and missing part is to fine-tune the decoder/prompt encoder on the video mode

@ronghanghu