ArliAI/GLM-4.5-Air-Derestricted
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, o_proj changed, in 26 of 47 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/2, Norms 0/98, MLP 494/2568, Attention 26/329, lm_head 0/1, other 0/2
12 derivatives checked
- mradermacher/GLM-4.5-Air-Derestricted-GGUF: Quantization
- bartowski/ArliAI_GLM-4.5-Air-Derestricted-GGUF: Quantization
- Koshkasa/ArliAI_GLM-4.5-Air-Derestricted_REAP50_Creative: LoRA-style (fewer layers)
- cyankiwi/GLM-4.5-Air-Derestricted-AWQ-4bit: Quantization
- DevQuasar/ArliAI.GLM-4.5-Air-Derestricted-GGUF: Quantization
- mradermacher/GLM-4.5-Air-Derestricted-i1-GGUF: Quantization
- ddh0/GLM-4.5-Air-Derestricted-GGUF: Same weights, converted
- noctrex/GLM-4.5-Air-Derestricted-MXFP4_MOE-GGUF: Quantization
- TMNE/ArliAI_GLM-4.5-Air-Derestricted-GGUF: Quantization
- nightmedia/GLM-4.5-Air-Derestricted-qx53g-mlx: Quantization
- ailexleon/GLM-4.5-Air-Derestricted-mlx-6Bit: Quantization
- Mawdistical-S1/Gaslit-106B: LoRA-style