nvidia/GLM-5.2-NVFP4
Quantization vs zai-org/GLM-5.2. nvidia/GLM-5.2-NVFP4 is a quantized version of the other (U8/BF16).
3 derivatives checked
- Jiunsong/SuperGLM-5.2-abliterated-NVFP4: Abliterated
- dealignai/GLM-5.2-NVFP4-UNCENSORED: Abliterated
- 0xSero/GLM-5.2-504B-Nvidia: Partial fine-tune (fewer layers)
7 other repos hold byte-for-byte identical weight files.