Flux: The Largest Open-Source Text-to-Image Model Outperforms Stable Diffusion
Black Forest Labs, known for creating Stable Diffusion, introduces Flux, a groundbreaking 12B parameter text-to-image model. Available on the fal platform, Flux offers superior aesthetics, drawing comparisons to Midjourney, and boasts three variations:
FLUX.1 [dev]: The foundation model, open-source with a non-commercial license, empowers community development.
FLUX.1 [schnell]: A distilled version prioritizing speed, operating up to 10 times faster than the base model, under the Apache 2 License.
FLUX.1 [pro]: A closed-source, API-accessible version for professional applications.
Unmatched Performance and Features
Flux leverages fal’s inference engine to deliver up to 2x faster processing speeds compared to eager torch, without compromising output quality. Key features include:
- Exceptional Image Quality: Generate stunning visuals at higher resolutions.
- Advanced Human Anatomy and Photorealism: Create lifelike images with accurate anatomical details.
- Improved Prompt Adherence: Obtain images that precisely match input prompts.
- Unmatched Speed: Particularly the Flux Schnell variation, caters to high-demand applications.
Black Forest Labs’ Flux signifies a major advancement in open-source text-to-image technology. Its impressive capabilities, combined with accessibility on the fal platform, empower both the AI community and professionals to explore new creative frontiers.