We provide four example notebooks that demonstrate how the micro_sam python library can be used for:

• Automatic instance segmentation with minimal effort in the quick_start.ipynb notebook.
• Automatic instance segmentation with more configuration options in the automatic_segmentation.ipynb notebook.
• Applying a Segment Anything Model (SAM) to many images and evaluating the segmentation quality against ground-truth segmentations in inference_and_evaluation.ipynb.
• Fine-tuning SAM on your custom data to obtain a better model for your segmentation tasks in sam_finetuning.ipynb.

If you have an environment with micro_sam you can directly run these notebooks. You can also run them on Kaggle Notebooks; we install the correct dependencies for it in each notebook.