cabooster/UDMT: UDMT v1.1.1

This release provides a frozen, citable snapshot of the UDMT codebase corresponding to the version used in the accompanying manuscript. Scope This release is intended solely for reproducibility purposes. The code snapshot matches the implementation used to generate the results reported in the manuscript. No further feature development or refactoring is intended for this release. Main contents Complete implementation of UDMT (Unsupervised Deep Multi-animal Tracking). Training, inference, and evaluation pipelines as described in the paper. Graphical user interface (GUI) for training, tracking, and visualization. Reproducibility notes The recommended software environment (OS, Python, PyTorch, CUDA) is described in the README. Python dependencies are specified in requirements.txt and requirements_custom.txt. Large datasets are not included in the repository and should be obtained as described in the paper and README. License The code is released under an Academic or Non-profit, Non-commercial Research Use Only license. See the LICENSE file for full terms and conditions. Citation If you use this code, please cite the accompanying paper as described in the Citation section of the README.