Recently, Black Forest Labs released the brand new FLUX.1Tools toolkit, designed to enhance the control and manipulability of its foundational text-to-image model, FLUX.1, enabling users to precisely modify and recreate both real and generated images.
FLUX.1Tools includes four main features:
FLUX.1Fill: This is an advanced image restoration and expansion tool that surpasses existing tools such as Ideogram2.0 and the open-source AlimamaCreative’s FLUX-Controlnet-Inpainting. FLUX.1Fill can edit and expand real and generated images based on text descriptions and binary masks provided by users, seamlessly integrating modifications into the images for a natural and smooth effect.
Tests have shown that the performance of FLUX.1Fill [pro] outperforms all other competitors, making it the most advanced image restoration model currently available. The open-source version, FLUX.1Fill [dev], also exceeds other proprietary solutions and offers higher inference efficiency.
FLUX.1Depth and FLUX.1Canny: Both tools utilize structural information from images to guide image generation, maintaining the original structure while modifying image textures.
FLUX.1Depth modifies images using depth information, while FLUX.1Canny uses edge detection information for guidance, making it more suitable for precise detail adjustments.
Evaluation results indicate that FLUX.1Depth outperforms proprietary models like Midjourney ReTexture, with FLUX.1Depth [pro] providing higher output diversity, while FLUX.1Depth [dev] offers more consistent results in depth perception tasks. For edge detection models, FLUX.1Canny [pro] delivers the best performance, followed by FLUX.1Canny [dev].
FLUX.1Redux: This is an adapter for all FLUX.1 foundational models, used for generating image variations. Given an input image, FLUX.1Redux can make subtle changes to create different versions, and it can even redesign the image style based on user-provided text descriptions.
Additionally, FLUX.1Redux supports image style adjustments through an API, requiring only an image and a text description from the user. This feature has been integrated into the latest FLUX1.1[pro] Ultra model, which combines input images and text prompts to create high-quality 4 million pixel outputs and supports flexible aspect ratios. Test results show that FLUX.1Redux achieves state-of-the-art performance in generating image variations.
FLUX.1Tools is available in both open-source and API versions.
FLUX.1Tools is offered in two versions: the open-source version (FLUX.1[dev]) and the API version (FLUX.1[pro]). Each tool's API version is released as a FLUX.1[pro] variant, while the open-source version is released as a guiding distilled FLUX.1[dev] variant, providing inference code and weights. Furthermore, Black Forest Labs has announced that the models released will be available through its partners: fal.ai, Replicate, Together.ai, Freepik, and krea.ai.
Black Forest Labs stated that the release of the FLUX.1Tools toolkit will inject new vitality into the Flux ecosystem, and they look forward to seeing community users create more exciting works with these new tools.
Model: https://huggingface.co/black-forest-labs/FLUX.1-Redux-dev
For more details: https://blackforestlabs.ai/flux-1-tools/
