grimjim/Magnolia-Mell-v1-12B

Partial fine-tune vs grimjim/Magnolia-v3-12B. Partial change: 87% of the weights changed (attention, mlp, norms), in 40 of 40 layers.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 27/81, MLP 117/120, Attention 156/160

3 derivatives checked

On the Hub · Downloads on Model Pulse