MOPO Labs presents what to our knowledge is the first quantized GLM-4.5 with a 2-bit MLX format quantization that allows the original 355gb model to work on a 192GB Mac using LM Studio and other MLX compatible tools. Note this is a research build intended to learn more about the differences between the GLM-4.5 quantized lower versus the GLM-4.5 Air model; If you discover good tips, kindly share with monir@mopo.life or on huggingface.
- Downloads last month
- 8
Model size
353B params
Tensor type
BF16
路
U32
路
F32
路
Hardware compatibility
Log In
to add your hardware
2-bit
Model tree for monirmamoun/GLM-4.5-MLX-2bit
Base model
zai-org/GLM-4.5