Text-to-Image
Diffusers
Safetensors
AXQReferenceDiffusionPipeline
axquant
axq
nvfp4
reference-storage
research-only
development
Instructions to use AutomatosX/AX-Image-2.1-Turbo-AXQ-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use AutomatosX/AX-Image-2.1-Turbo-AXQ-NVFP4 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("AutomatosX/AX-Image-2.1-Turbo-AXQ-NVFP4", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
Download Notice from AutomatosX/AX-Image-2.1-Turbo-AXQ-NVFP4: direct link, hf CLI and curl.
- Browser
- Download file 386 Bytes
-
https://huggingface.co/AutomatosX/AX-Image-2.1-Turbo-AXQ-NVFP4/resolve/main/Notice
- Command line
-
hf download hf://AutomatosX/AX-Image-2.1-Turbo-AXQ-NVFP4/Notice
-
curl -L -o Notice https://huggingface.co/AutomatosX/AX-Image-2.1-Turbo-AXQ-NVFP4/resolve/main/Notice
386 Bytes
| Qwen is licensed under the Qwen RESEARCH LICENSE AGREEMENT, Copyright (c) 2026 Hangzhou Tongyi Laboratory Technology Co., Ltd. All Rights Reserved. | |
| AXQuant changes (2026-10-14): selected transformer matrices are packed into NVFP4 reference storage. The pipeline entrypoint requires explicit BF16 reconstruction. No native low-precision execution or quality certification is claimed. | |