PruhaNLP/USER2-1C-code

Partial fine-tune vs deepvk/USER2-base. Partial change: 74% of the weights changed (attention, embeddings, mlp, norms), in 22 of 22 layers.

Tensors changed: Embeddings 1/2, Norms 44/44, Attention 44/44, MLP 44/44

1 derivatives checked

On the Hub · Downloads on Model Pulse