Are there any detailed instructions to use pretrained models with custom dataset?

[Mahesha999]

I want to know if there any instructions to use the MOT pretrained models to track on custom dataset (assuming the data set objects are similar to what we have in MOT datasets). I checked track.py and it has MOT dataset specific hardcoded lines (for example line 340 to set track_buffer and lint 347 to set track_high_thresh).

While I can go through all necessary script and make necessary changes to use pretrained models to try tracking on custom dataset, can someone let me know if there are any written instructions for this? Or if someone has tried it earlier? Also are there any more detailed instructions to train the model on custom dataset (apart from the train section in readme)?