[Feature] Please support Flex.1-alpha

[nitinmukesh]

Checklist

• [x] 1. If the issue you raised is not a feature but a question, please raise a discussion at https://github.com/mit-han-lab/nunchaku/discussions/new/choose. Otherwise, it will be closed.
• [x] 2. I will do my best to describe the issue in English.

Motivation

Flex.1 alpha is a pre-trained base 8 billion parameter rectified flow transformer capable of generating images from text descriptions. It has a similar architecture to FLUX.1-dev, but with fewer double transformer blocks (8 vs 19). It began as a finetune of FLUX.1-schnell which allows the model to retain the Apache 2.0 license. A guidance embedder has been trained for it so that it no longer requires CFG to generate images.

Model Specs 8 billion parameters Guidance embedder True CFG capable Fine tunable OSI compliant license (Apache 2.0) 512 token length input

Usage The model can be used almost identically to FLUX.1-dev and will work out of the box with most inference engines that support that. (Diffusers, ComfyUI etc.)

For ComfyUI, there is an all in one file called Flex.1-alpha.safetensors. Put this in your checkpoints folder and use like you would FLUX.1-dev.

Related resources

Flex.1-alpha https://huggingface.co/ostris/Flex.1-alpha

Flex.1-alpha 512 Redux Adapter https://huggingface.co/ostris/Flex.1-alpha-Redux

[nitinmukesh]

Sorry I searched before posting without result (title was incorrect in earlier request) so posted new

Please could you provide script. https://github.com/mit-han-lab/nunchaku/issues/228#issuecomment-2759529167

[lmxyy]

We will release a node for quantizing FLUX-like models. Hopefully, you can quantize the model yourself at that time.

[github-actions[bot]]

This issue has been automatically closed due to 30-day inactivity. Please feel free to reopen it if needed.