Together AI announced the integration of three advanced image creation models from Black Forest Labs’ FLUX Tools suite, with the goal of providing developers with improved control over image composition, depth, and style. This development marks a significant step forward in AI-based image creation and provides innovative capabilities for developers and artists alike.
Introducing a new model
The newly launched models, Canny, Depth, and Redux, are designed to encourage a variety of creative possibilities by allowing users to iterate on existing images. The Canny model focuses on precise compositional control through edge detection, while the Depth model uses depth mapping to maintain spatial relationships. Redux, on the other hand, creates subtle variations of images without text prompts, making it ideal for polishing existing visuals.
Advanced image control via ControlNet
At the core of these models is ControlNet, a neural network architecture developed at Stanford. This technique allows precise image generation control by using additional input images as a guide, providing a structural blueprint for the model to follow. This results in more predictable output and fewer iterations required to achieve the desired image.
Real-world applications and tools
These models are available through Together AI’s API, allowing developers to seamlessly integrate them into their projects. The availability of open source example apps such as Blinkshot and LogoCreator further demonstrates the practical application of these models. Developers can also take advantage of free endpoints for the FLUX schnell model until the end of the year, giving them the opportunity to explore these tools without financial burden.
For developers interested in experimenting with these models, Together AI provides a Python SDK for easy integration. This SDK allows you to quickly and efficiently integrate FLUX models into a variety of applications, enhancing your creative process with cutting-edge technology.
For more information about the model, see the Together AI blog.
Image source: Shutterstock