djuna/MS-Nudion-22B

Full fine-tune vs knifeayumu/Cydonia-v1.3-Magnum-v4-22B. Fine-tune: 100% of the weights changed, across 56 of 56 layers.

Tensors changed: lm_head 1/1, Embeddings 1/1, Norms 62/113, MLP 168/168, Attention 224/224

2 derivatives checked

• mradermacher/MS-Nudion-22B-i1-GGUF: Quantization
• mradermacher/MS-Nudion-22B-GGUF: Quantization

On the Hub · Downloads on Model Pulse