Instructions to use KevinHuang/OmniX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use KevinHuang/OmniX with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("KevinHuang/OmniX", 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
About the license
Hi Kevin, the base OmniX model is licensed under a non-commercial license, but OmniX itself is licensed under Apache 2.0. What does this mean?
Hi, thanks for the question. To clarify, the OmniX codebase/framework is licensed under Apache License 2.0, while the upstream/base model (Flux.1-dev) remains subject to its original non-commercial license. We only provide LoRA/adapters and do not redistribute the base model weights, so use of the base model continues to be governed separately by its original licensing terms.
Hi, thanks for the question. To clarify, the OmniX codebase/framework is licensed under Apache License 2.0, while the upstream/base model (Flux.1-dev) remains subject to its original non-commercial license. We only provide LoRA/adapters and do not redistribute the base model weights, so use of the base model continues to be governed separately by its original licensing terms.
Thank you very much for your response, Dr. Huang. Do you think it’s possible to make it more commercially viable using base models such as Flux Schnell, Klein-4b, Z-Image, or Qwen-Image? I want to generate panoramas for my application and am currently using HY-PANO 2.0 from HY-WORLD 2.0, but it requires too many resources :'(.