Instructions to use abenzerps/Qwen-Image-2.1-Turbo-Quantized with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use abenzerps/Qwen-Image-2.1-Turbo-Quantized with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download abenzerps/Qwen-Image-2.1-Turbo-Quantized --local-dir Qwen-Image-2.1-Turbo-Quantized
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Unsloth Desktop Support
I know this model is geared primarily towards supporting ComfyUI. Are there any thoughts on supporting Unsloth Desktop? I have tried installing it, but it seems to fail the diffusion step when being ran becuase of some incompatibility with the loaded model and the base model:
abenzerps/Qwen-Image-2.1-Uncensored-GGUF
qwen-image-2.1-UC-Q8_0.gguf
mismatch: 4096 vs 8192 for the GCUF conversion/quantization
Any support or debug steps or anything I could do to help would be appreciated.
hey @Ginger0913 ,
why it fails:
these GGUFs are packaged specifically for ComfyUI. Unsloth Desktop runs on stable-diffusion.cpp, which expects a different internal tensor structure—that's what triggers the 4096 vs 8192 mismatch.
current workaround:
for now, the model runs smoothly in ComfyUI using Unet Loader (GGUF) with the official Qwen3-VL text encoder and VAE.
have a great one!