DaMiTian/smolvla-aloha-bimanual

Partial fine-tune vs lerobot/smolvla_base. Partial change: 23% of the weights changed (attention, mlp, other), in 0 of 16 layers.

Tensors changed: other 6/7, MLP 52/148, Attention 64/224, Norms 0/116, lm_head 0/1, Embeddings 0/4

On the Hub · Downloads on Model Pulse