ArliAI/GLM-4.6-Derestricted-v3
LoRA-style vs zai-org/GLM-4.6. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate, o_proj changed, in 32 of 93 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/376, MLP 241/1970, Attention 32/651, lm_head 0/1, other 0/1
5 derivatives checked
- bartowski/ArliAI_GLM-4.6-Derestricted-GGUF: Quantization
- petra312/GLM-4.6-Derestricted-v3-NF4: Quantization
- mradermacher/GLM-4.6-Derestricted-i1-GGUF: Quantization
- mradermacher/GLM-4.6-Derestricted-GGUF: Quantization
- mradermacher/GLM-4.6-Derestricted-v3-i1-GGUF: Not comparable yet