license: apache-2.0
base_model: huihui-ai/Huihui-Qwen3-VL-32B-Instruct-abliterated
tags:
- awq
- quantized
- qwen3-vl
- abliterated
- uncensored
- vision-language
- not-for-all-audiences
pipeline_tag: image-text-to-text
Huihui-Qwen3-VL-32B-Instruct-abliterated AWQ
AWQ (4-bit) quantization of huihui-ai/Huihui-Qwen3-VL-32B-Instruct-abliterated, which is an abliterated (uncensored) version of Qwen/Qwen3-VL-32B-Instruct.
A vision-language model: it reads images and answers in text. Useful for detailed image captioning and prompt writing. The AWQ version needs about 20 GB of VRAM and runs fast with vLLM.
Usage (vLLM)
vllm serve draganjankovic1975/Huihui-Qwen3-VL-32B-Instruct-abliterated-AWQ --max-model-len 8192
Warning
This model is abliterated: safety refusals were removed. It may produce sensitive or inappropriate content. Not for public-facing applications, not for minors. You are responsible for how you use it and what it generates.
Credits and license
- Original model: Qwen team – Qwen3-VL-32B-Instruct
- Abliteration: huihui-ai
- AWQ quantization: draganjankovic1975
Licensed under the Apache License 2.0 (see LICENSE), same as the original model.