knifeayumu/Cydonia-v1.2-v1.3-Magnum-v4-22B

Full fine-tune vs anthracite-org/magnum-v4-22b. Fine-tune: 100% of the weights changed, across 56 of 56 layers.

Tensors changed: Embeddings 1/1, Norms 63/113, MLP 168/168, Attention 224/224, lm_head 1/1

1 derivatives checked

On the Hub · Downloads on Model Pulse