1 results for tag "tribev2-brain-encoding"
Runs Meta TRIBE v2 — a multimodal foundation model that predicts fMRI brain responses on the fsaverage5 cortical surface (~20k vertices) from video, audio, and text stimuli, combining LLaMA 3.2 + V-JEPA2 + Wav2Vec-BERT encoders via tribev2.TribeModel.from_pretrained("facebook/tribev2") with nilearn/PyVista visualization.