win10/Qwen3.8-27B-EXP-EVE-v1
LoRA-style vs Qwen/Qwen3.8-27B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, linear_fc1, linear_fc2, o_proj, out_proj, proj, q_proj, qkv, up_proj, v_proj changed, in 64 of 64 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/325, Attention 155/560, MLP 246/307, Embeddings 0/5, lm_head 0/1, other 0/1
2 derivatives checked
- mradermacher/Qwen3.8-27B-EXP-EVE-v1-i1-GGUF: Quantization
- mradermacher/Qwen3.8-27B-EXP-EVE-v1-GGUF: Quantization