Vortex5/MN-12B-Azure-Veil
Partial fine-tune vs SicariusSicariiStuff/Impish_Nemo_12B. Partial change: 89% of the weights changed (attention, embeddings, lm_head, mlp, norms), in 35 of 40 layers.
Tensors changed: Embeddings 1/1, Norms 71/81, MLP 105/120, Attention 140/160, lm_head 1/1
2 derivatives checked
- mradermacher/MN-12B-Azure-Veil-GGUF: Quantization
- mradermacher/MN-12B-Azure-Veil-i1-GGUF: Quantization