prithivMLmods/Messier-Opus-14B-Elite7
Partial fine-tune vs prithivMLmods/Equuleus-Opus-14B-Exp. Partial change: 88% of the weights changed (attention, mlp, norms), in 48 of 48 layers.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 6/97, MLP 141/144, Attention 201/336
3 derivatives checked
- mradermacher/Messier-Opus-14B-Elite7-i1-GGUF: Quantization
- mradermacher/Messier-Opus-14B-Elite7-GGUF: Quantization
- prithivMLmods/Lacerta-Opus-14B-Elite8: Full fine-tune