pacozaa/mistral-unsloth-chatml-first

Partial fine-tune vs unsloth/mistral-7b-bnb-4bit. Partial change: 3% of the weights changed (attention, mlp), in 32 of 32 layers.

Tensors changed: MLP 203/491, Attention 277/658, lm_head 0/1, Embeddings 0/1, Norms 0/65

On the Hub · Downloads on Model Pulse