Releases: weigertlab/trackastra
Releases · weigertlab/trackastra
0.5.2
0.5.1
0.5.0
What's Changed
- SAM2.1 Bacteria - Pre-trained features model release by @C-Achard in #39
- Update CLI by @bentaculum in #47
- Set pythonic default model download dir by @bentaculum in #49
New Contributors
Full Changelog: 0.4.2...0.5.0
0.4.2
What's Changed
- Support empty windows+sequences in prediction by @bentaculum in #42
- Make installing dependencies for training optional by @bentaculum in #43
Full Changelog: 0.4.1...0.4.2
0.4.1
Update to GEFF v1 API
0.4.0
What's Changed
- Breaking change to python inference API:
Trackastra.tracknow takes care of relabeling the segmentation masks to match IDs in the solution graph, which before was done only when converting to CTC format.
Here is a minimal example of the new API:# load some test data images and masks imgs, masks = example_data_bacteria() # Load a pretrained model model = Trackastra.from_pretrained("general_2d", device=device) # or from a local folder # model = Trackastra.from_folder('path/my_model_folder/', device=device) # Track the cells track_graph, masks_tracked = model.track(imgs, masks, mode="greedy") - Add GEFF file format support in inference by @bentaculum in #34
- Support inference on large arrays with dask by @maweigert in #33
- Add batched inference by @bentaculum in #35
- Support intermediate empty frames in inference by @JojoDevel in #36
Full Changelog: 0.3.2...0.4.0
0.3.2
Full Changelog: 0.3.1...0.3.2
0.3.1
What's Changed
- Update setup.cfg file to install
humanizepackage by @anwai98 in #30 - Add basic train test
- Add python 3.13 compatibility
Full Changelog: 0.3.0...0.3.1
0.3.0
What's Changed
- Remove verbosity argument in LRScheduler by @anwai98 in #27
- Update training, release new models by @bentaculum in #29
New Contributors
Full Changelog: 0.2.4...0.3.0
0.2.4
Full Changelog: 0.2.3...0.2.4