mlx-community/DeepSeek-R1-Distill-Llama-8B-4bit

Quantization vs deepseek-ai/DeepSeek-R1-Distill-Llama-8B. mlx-community/DeepSeek-R1-Distill-Llama-8B-4bit is a quantized version of the other (U32/F16).

4 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse