Emanon14/NONAMEmix-Vpred_v2

Full fine-tune vs Minthy/RouWei-0.7. Fine-tune: 97% of the weights changed, across 11 of 12 layers.

Tensors changed: other 650/662, Embeddings 568/590, Attention 716/716, Norms 546/546

On the Hub · Downloads on Model Pulse