Latent dynamics (IDM/FDM) · libero_90 · seed 1000
Action-free visual dynamics pretrain: frozen SmolVLA vision features → IDM → FDM.
Trained on RGB from nvidia/libero_90 only (no action / state / reward labels).
Files:
idm_fdm.pt — model weights + config dictidm_fdm.json — training hyperparameters
Used as frozen teacher for video-latent few-shot SmolVLA fine-tuning
(--video.checkpoint).
