fix(grit): pool object features with full per-image box list

fix(grit): split logits per image for batched inference

fix(grit): enable multi-image ROIHeads inference path

feat: batch GrIT inference for 4 slow dimensions

Revise video generation models information

Update README.md

Update README.md

Update I2V project links and details in README

Update I2V model details in README.md

Update README with new model details

'[update] Human_Anatomy'

Update video model links and details in README

Add details of video generation models to README

Remove 'ToMovie 2.0' entry from README

Remove redundant link to Tomovie website