gen-robot/openvla-7b-rlvla-sft_16k

Full fine-tune vs gen-robot/openvla-7b-rlvla-warmup. Fine-tune: 97% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 2/7, Norms 0/275, MLP 188/304, Attention 226/343, other 1/52, lm_head 1/1

On the Hub · Downloads on Model Pulse