FLUX.1 Canny/Depth: Precise Image Transformation

Structural Conditioning with FLUX.1 Canny/Depth: Precise Image Transformation

BlackForestLabs expands the boundaries of image editing with FLUX.1 Canny and FLUX.1 Depth, tools designed to offer precise structural conditioning. These models enable controlled transformations using canny edges and depth detection, ensuring core composition remains intact.

Click to try this feature

What is Structural Conditioning?

Structural conditioning uses edge or depth maps to guide image modifications. This approach maintains the original image’s structure while enabling detailed, text-guided edits. It is especially effective for retexturing, allowing changes to textures without altering the foundational structure.

Benchmark-Topping Performance

• FLUX.1 Depth:
  • Outperforms proprietary models like Midjourney ReTexture.
  • The [pro] version delivers diverse outputs, while the [dev] version ensures consistency in depth-aware tasks.
• FLUX.1 Canny:
  • Recognized as the best in class, with the [pro] version leading benchmarks, followed closely by the [dev] version.

Versions and Accessibility

• FLUX.1 Canny / Depth [dev]:
  • Available under the Flux Dev License.
  • Full and LoRA model weights on [Hugging Face].
  • Inference code: Available on [GitHub].
• FLUX.1 Canny / Depth [pro]:
  • Offered via the BFL API.

Harness the power of structural conditioning with FLUX.1 Canny and Depth for precise, guided image transformations.

How to try

This model is now available on our platform. You only need to log in and subscribe to use this model and other latest model features. Click to try this feature