Finetune on custom dataset

Hi, thank you for this splendid work.

Could you explain how can I finetune sam3 on my own video dataset and how should I annotate my dataset?

Thanks!

No specific file, notebook, annotation format, or finetuning entry point is identified in the issue. Review the repository's finetuning implementation and example notebooks first, then define the supported custom video-dataset workflow and annotation guidance; done means the explanation is complete and reproducible.

Written by the indexing model from the issue text.