metadata
library_name: mlx
tags:
- mlx
- oq
- quantized
This quantization was uploaded on 2026-06-29 and replaces a previous version. If you downloaded this model before this date, please re-download for the updated weights.
Mistral-7B-Instruct-v0.2-oQ4
This model was quantized using oQ (oMLX v0.4.3) mixed-precision quantization.
Quantization details
- Model type: mistral
- Bits: 4
- Group size: 64
- Format: MLX safetensors