samuelcardillo/Carnice-Qwen3.6-MoE-35B-A3B-GGUF
LoRA-style vs Qwen/Qwen3.6-35B-A3B. samuelcardillo/Carnice-Qwen3.6-MoE-35B-A3B-GGUF is a GGUF (F16) of different weights. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 10 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/101, MLP 0/200, Attention 40/40, lm_head 0/1