Corrected text
[change: seg-v4] ML Models and Trajectory Segmentation: segmentation model v4 (alpha = 0 for immobile segments), benchmark
[change: no-immobile-rule] ML Models: no immobile rule
[change: c15-strict-outputs] ML Models troubleshooting: per-frame output length error
[change: model-contributions] ML Models: contributing a model
[change: ml-models-wording] ML Models: inference runs on CPU or GPU
[change: benchmark-links] ML Models: link to the online benchmark tool
small corrections
[change: seg-model-rules] Document Model segments for custom models, the default model's immobile rule, and verbatim custom α
[change: seg-inference] Bundled segmentation model: TTA always on inside the model, parallel inference, segment rules
[change: analysis-speed] Default analysis model: compute grows with length, trajectories run in parallel
[change: analysis-speed] Analysis model compute and parallel inference
[change: analysis-v1] Describe the FCI-type default analysis model
upd to v0.1.0 code (separation frame and segment)
Delete all model_type refs
change model_type to diffusive_state
Update ML-Models.md
added new utils links
correction pass
Update segmentation + others
continue upds
finish draft landing page