About ToM.
Where in the code can I see the ToM (algorithm2) mentioned in the paper? I would like to see the detailed code of the ToM algorithm using CoTracker.
And I wonder whether the segments using PySceneDetect in ToM contain semantic sub-goals in each video task.
However, I cannot find the code. I would appreciate your help.
Will release the code soon this week!
@jwyang Thanks for uploading the quick training code. I am more interested in ToM than overall training model. Which part should I check to test the ToM algorithm mentioned in the paper? (algorithm2)
@hahamini , please follow this part to try the som_tom (Alg2) demo:
https://github.com/microsoft/Magma?tab=readme-ov-file#som-and-tom-generation