What is the process for obtaining multiple mask outputs using either the model.predict() or the model.mask_decoder() method when performing instance segmentation finetuning?

facebookresearch / segment-anything

The repository provides code for running inference with the SegmentAnything Model (SAM), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

Apache License 2.0

47.39k stars 5.61k forks source link

What is the process for obtaining multiple mask outputs using either the model.predict() or the model.mask_decoder() method when performing instance segmentation finetuning? #532

Open Soham-Objectways opened 1 year ago

Soham-Objectways commented 1 year ago

I am seeking guidance on performing SAM fine-tuning for multi-mask instance segmentation. Additionally, during the inference stage, I wish to utilize the model.predict function to predict multiple instance segmentation masks without any prompts. This is because the actual inference images may not have any bounding boxes available.

cristian-cmyk4 commented 8 months ago

Hi, could you find a solution? I have the same problem