v1.1.0 🌈

Version 1.1.0

micro_sam is a tool for microscopy image segmentation based on Segment Anything model.

This version introduces several improvements:

• Compatibility with napari >= 0.5
• CLI for automatic instance segmentation
• Initial support for parameter efficient finetuning and automatic 2d/3d semantic segmentation (not available via the napari plugin yet, but part of the python library)
• And several other minor improvements and bugfixes

Changes

• Bump to version 1.1 @constantinpape (#745)
• Changes for release 1.1 @constantinpape (#744)
• Add materials for hands-on workshops @anwai98 (#742)
• Disable thread workers @constantinpape (#739)
• Fix link to the finetuning notebook in doc @anwai98 (#743)
• Updates to the image series annotator @constantinpape (#738)
• Expose is_seg_dataset argument in sam dataset @anwai98 (#736)
• Add CLI for benchmarking datasets on SAM models @anwai98 (#728)
• Add dry run for loaders to check for valid instances @anwai98 (#705)
• Minor update to ignore warnings in train_sam functionality @anwai98 (#722)
• Update docs - add hints on starting napari @anwai98 (#724)
• Add recommendations in finetuning notebook @anwai98 (#720)
• Expose kwargs to allow changing peft module arguments @anwai98 (#718)
• Add selective peft methods @anwai98 (#708)
• Update dropout defaults for FacT finetuning method @anwai98 (#714)
• Overwrite keybindings of prompt layers @constantinpape (#711)
• Extend iterative prompt generators to return prompts for 3d @anwai98 (#692)
• Napari updates @constantinpape (#710)
• Revert "Fix clearing shape layer" @constantinpape (#709)
• Fix clearing shape layer @constantinpape (#696)
• Add section in FAQ for passing custom model weights @anwai98 (#707)
• Add tests for automatic segmentation function @anwai98 (#702)
• Add automatic segmentation cli @anwai98 (#699)
• Update yaml files - pin napari < 0.5 @anwai98 (#694)
• Fix livecell top-level finetuning script - adapt with new peft changes @anwai98 (#691)
• Add nifty pin to yaml files @anwai98 (#688)
• Minor update to move edge-related attributes to border @anwai98 (#689)
• Minor update to peft kwargs in SAM wrappers @anwai98 (#686)
• Add FaCT Finetuning for SAM @anwai98 (#682)
• Update the python versions in test matrix @constantinpape (#685)
• Add LoRA support to LIVECell finetuning scripts @anwai98 (#681)
• Minor updates to resource efficient finetuning experiments @anwai98 (#665)
• Minor fix to label transform function in finetuning notebook @anwai98 (#677)
• Add livecell lora plots @anwai98 (#674)
• Fix issue in AIS precomputation and extend training functionality @constantinpape (#673)
• Fix issue with dtypes that leads to integer overflow in committing @constantinpape (#671)
• Add plotting scripts for AIS experiments @anwai98 (#666)
• Remove disconnected objects after AMG @constantinpape (#668)
• Update docstrings to document recent functionalities @anwai98 (#664)
• Add AIS benchmarking scripts @anwai98 (#657)
• Minor update to input checking for training functionality @anwai98 (#663)
• Add min-size to training and fix other issues @constantinpape (#658)
• Add weighting to dice loss in semantic trainer @anwai98 (#656)
• Add timing outputs @constantinpape (#655)
• Minor refactor to LoRA model initialization in get_trainable_sam_model @anwai98 (#654)
• Refactor resource efficient finetuning scripts @anwai98 (#653)
• Fix bug in precompute for 3d data @constantinpape (#649)
• Api cleanup @constantinpape (#648)
• Minor fix to trainable sam model functionality @anwai98 (#646)
• Add simple 3d wrapper and enable freezing the encoder in sam 3d wrapper @constantinpape (#645)
• Add SemanticSam3dLogger @anwai98 (#643)
• 636 enhance 3d image processing based on ma sam @lufre1 (#639)
• Minor fix to loading incompatible layers @anwai98 (#641)
• Minor update to loading models @anwai98 (#640)
• Add SemanticSamTrainer @anwai98 (#637)
• Add LoRA Implementation @anwai98 (#611)
• Minor update to default rescaling params in resizerawtrafo @anwai98 (#635)
• Add SimpleSamTrainer @anwai98 (#631)
• Add mentions for annotating 3D RGB volumes @anwai98 (#629)
• Remove device input for TrainableSAM @anwai98 (#632)
• Add MedSAM Reimplementation @anwai98 (#612)
• Add PanNuke specialist finetuning @anwai98 (#622)
• Upstream dev with master @anwai98 (#624)
• Add progressbar to pooch download @constantinpape (#623)
• Update docs @constantinpape (#621)