Lambent/Mira-v1.21-27B-rlvr
LoRA-style vs Lambent/Mira-v1.20-27B-dpo. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 62 of 62 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/4, Norms 0/484, MLP 182/294, Attention 234/464, other 0/1
5 derivatives checked
- mradermacher/Mira-v1.21-27B-rlvr-i1-GGUF: Quantization (fewer layers)
- mradermacher/Mira-v1.21-27B-rlvr-GGUF: Quantization (fewer layers)
- virtuous7373/Lambent-Mira-Calliope: Full fine-tune
- Lambent/Mira-v1.25-27B-Wave: Full fine-tune
- virtuous7373/Mira-Demeter: Full fine-tune