DAMO-NLP-SG/VideoLLaMA3-2B

Full fine-tune vs DAMO-NLP-SG/VideoLLaMA3-2B-Image. Fine-tune: 100% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 3/3, Norms 143/167, MLP 191/192, Attention 412/412, other 4/4

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse