Finetune SAM for instance segmentation to output segmenatation masks along with label names

Hello,

It seems the SAM model from the transformers library does not directly support classification tasks out of the box, as it is primarily designed for generating segmentation masks.

Any way or resources to finetune SAM for instance segmentation to output segmenatation masks along with label names please?

Thanks

The issue names no file, test, or entry point. Start by reviewing the repository’s example notebooks and inference code, then clarify the intended finetuning and label-output scope with maintainers; done requires an agreed implementation target and a demonstrated instance-segmentation result.

Written by the indexing model from the issue text.