Lambent/Mira-v1.13-27B-custom-heretic-ties
Full fine-tune vs Lambent/Mira-v1.12-Ties-27B. Fine-tune: 98% of the weights changed, across 62 of 62 layers.
Tensors changed: Embeddings 1/4, Norms 162/484, MLP 186/294, Attention 248/464, other 1/1
3 derivatives checked
- mradermacher/Mira-v1.13-27B-custom-heretic-ties-i1-GGUF: Quantization (fewer layers)
- mradermacher/Mira-v1.13-27B-custom-heretic-ties-GGUF: Quantization (fewer layers)
- Lambent/Mira-v1.14-27B: LoRA-style