---
library_name: transformers
license: apache-2.0
pipeline_tag: image-text-to-text
base_model:
- Qwen/Qwen3.5-27B
tags:
- abliterated
- uncensored
- mlx
---

# Huihui-Qwen3.5-27B-abliterated-8bit-MLX

This is an 8-bit quantized MLX version of [huihui-ai/Qwen3.5-27B-abliterated](https://huggingface.co/huihui-ai/Qwen3.5-27B-abliterated), a vision-language model based on Qwen3.5-27B with abliteration applied to remove refusal behavior.

## Usage

```bash
mlx-vlm generate --model AITRADER/Huihui-Qwen3.5-27B-abliterated-8bit-MLX --prompt "Describe this image." --image <path_to_image>
```