Black Forest Labs released four new artificial intelligence (AI) tools for its base text-to-image Flux.1 AI model last week. These four AI tools are based on separate models designed to execute specific image editing tasks within the image generator. The company claims that these tools will offer granular control over the output images and let users preserve the key elements while experimenting with different styles. The editing tools are available separately as developer models in open access and pro models.
Black Forest Labs Releases Image Editing Tools for Flux.1
In a blog post, the AI firm detailed the four new image editing tools for Flux.1 AI model. Developers can open-access the four tools in separate models within the dev model series, whereas users will get access to the full version via BFL API.
Today, we are excited to release FLUX.1 Tools, a suite of models designed to add control and steerability to our base text-to-image model FLUX.1, enabling the modification and re-creation of real and generated images. Learn more in our blogpost: https://t.co/J5Bc8fVGEc pic.twitter.com/7lEl74XYV4
— Black Forest Labs (@bfl_ml) November 21, 2024
The Flux.1 Fill is an inpainting and outpainting tool that can edit the details within an image or expand the boundaries of an image using text prompts and a binary mask. Based on internal testing, the company claimed that the pro version of the tool outperforms competing tools such as Ideogram 2.0. The developer version of the tool is available under the Flux Dev License and can be found on Hugging Face and GitHub. The pro version can be accessed via the BFL API.
Flux.1 Depth and Flux.1 Canny tools let users perform structural conditioning of output during image transformations. The Depth tool preserves the generated image's structure through depth maps and keeps it intact while users make a text-guide edit. Similarly, the Canny tool preserves the structure by accessing the output's canny edges. These are helpful during retexturing-based edits.
The company claimed that the tools outperform similar tools offered by competitors such as Midjourney and InstantX. The full version offers maximum performance whereas the Low-rank adaptation (LoRA) version for developers allows for easier deployment. It can be found here.
Finally, the Flux.1 Redux allows users to generate image variations based on an input image. Black Forest Labs claims the tool can reproduce the image with slight variation, which can be refined further. It also allows image restyling via prompts. Also, this tool is supported by the Flux1.1 [pro] Ultra, the company's flagship image generation model. The model weights can be found here.
All of the AI models will also be available via third-party platforms such as Fal.ai, Replicate, Together.ai, Freepik, and Krea.ai.
