ConicCat/GLM-4.5-Architect-106B-A12B
LoRA-style vs zai-org/GLM-4.5-Air. 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 46 of 47 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/2, Norms 0/98, MLP 2415/2568, Attention 184/329, lm_head 0/1, other 0/2
3 derivatives checked
- mradermacher/GLM-4.5-Architect-106B-A12B-i1-GGUF: Quantization
- mradermacher/GLM-4.5-Architect-106B-A12B-GGUF: Quantization
- flukethoughts/GLM-4.5-Architect-106B-A12B-NVFP4: Quantization