ReadyArt/Omega-Evolution-27B-v2.1

LoRA-style vs ConicCat/Qwen3.5-27B-Writer-V2. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, k_proj, o_proj, out_proj, q_proj, up_proj, v_proj changed, in 64 of 64 layers; norms, biases and embeddings are untouched.

Tensors changed: lm_head 0/1, Embeddings 0/4, Norms 0/319, Attention 304/556, MLP 192/304

2 derivatives checked

On the Hub · Downloads on Model Pulse