mlx-community/DaisyRose-27B-4bit

This model was converted to MLX format (4-bit affine quantization, group_size=64) from RadixArk/DaisyRose-27B (built on Qwen/Qwen3.8-27B) using mlx-vlm version 0.8.0.

It preserves the full multimodal vision tower (image-text-to-text) and can be used with both mlx-vlm (text + image/video) and mlx-lm (text-only).

Refer to the original model card for architecture details, benchmarks, and license terms (Apache-2.0).

Usage

With mlx-vlm (Multimodal & Text)

pip install -U mlx-vlm
python -m mlx_vlm generate \
  --model mlx-community/DaisyRose-27B-4bit \
  --max-tokens 256 \
  --prompt "Describe this image." \
  --image <path_to_image>

With mlx-lm (Text-Only / Local Server)

pip install -U mlx-lm
mlx_lm.generate \
  --model mlx-community/DaisyRose-27B-4bit \
  --max-tokens 256 \
  --prompt "Write a quick bash script to check open ports."
Downloads last month
-
Safetensors
Model size
27B params
Tensor type
U32
ยท
BF16
ยท
MLX
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for mlx-community/DaisyRose-27B-4bit

Base model

Qwen/Qwen3.8-27B
Quantized
(21)
this model