Nex-N2-mini-MLX-VLM-8bit

Native MLX-VLM 8-bit quantized version of .

Summary

• Base model:
• Format: native MLX / MLX-VLM
• Quantization: 8-bit MLX-VLM quantization
• Group size: 64
• Vision: supported
• MTP: not included
• Target runtime: MLX-VLM / oMLX / Apple Silicon

This version is intended as a direct compatible release.

Quick test

python3 -m mlx_vlm.generate \
  --model joowon-jang/Nex-N2-mini-MLX-VLM-8bit \
  --image /path/to/image.jpg \
  --prompt "Describe this image in one sentence." \
  --max-tokens 128 \
  --temp 0.0

MTP variant

For oMLX Native MTP speculative decoding, use:

License

Apache-2.0, following the base model license.