Skip to content

TilingFluxImg2ImgPipeline

You need to perform setup steps to use Hugging Face nodes

This guide will walk you through setting up a Hugging Face account, creating an access token, and installing the required models to make this node fully functional.

What is it?

The TilingFluxImg2ImgPipeline node is designed to transform images using the FLUX.1 model with a tiling approach. It allows for image-to-image transformations based on text prompts, utilizing Hugging Face's diffusers library to apply changes across image tiles for efficient processing.

When would I use it?

Use this node when you need to:

  • Apply image-to-image transformations using text prompts
  • Process large images by dividing them into smaller tiles
  • Utilize advanced models for creative image manipulation
  • Experiment with different tiling strategies for image processing

How to use it

Basic Setup

  1. Add the TilingFluxImg2ImgPipeline node to your workflow
  2. Connect an image source to the "input_image" parameter
  3. Set the desired model in the "model" parameter
  4. Provide a text prompt in the "prompt" parameter
  5. Optionally configure additional parameters such as "tile_size", "tile_overlap", and "tile_strategy"
  6. Run the node to transform the image

Parameters

  • model: The FLUX.1 model to use for image transformation (default: "black-forest-labs/FLUX.1-schnell")
  • input_image: The image to be transformed (ImageArtifact)
  • prompt: The primary text prompt for image transformation
  • prompt_2: An optional secondary prompt (defaults to the primary prompt if not provided)
  • negative_prompt: An optional negative prompt to influence image transformation
  • negative_prompt_2: An optional secondary negative prompt (defaults to the primary negative prompt if not provided)
  • true_cfg_scale: A float value to adjust the configuration scale (default: 1.0)
  • num_inference_steps: The number of inference steps for image transformation (default: 16)
  • strength: A float value indicating the transformation strength (default: 0.3)
  • guidance_scale: A float value to adjust the guidance scale (default: 3.5)
  • seed: An optional integer seed for random number generation (default: random)
  • scale: A float value for scaling the image (default: 1.0)
  • tile_size: The size of each tile in pixels (default: 1024)
  • tile_overlap: The overlap between tiles in pixels (default: 64)
  • tile_strategy: The strategy for processing tiles (default: "linear"). Options include:
    • "linear"
    • "chess"
    • "random"
    • "inward"
    • "outward"
  • output_image: The transformed image as an ImageArtifact
  • logs: Logs of the image transformation process

Outputs

  • output_image: The final transformed image as an ImageArtifact
  • logs: A string containing logs of the image transformation process

Important Notes

  • The node requires a valid Hugging Face API token set as the environment variable HUGGINGFACE_HUB_ACCESS_TOKEN
  • Only specific FLUX.1 models are supported: "black-forest-labs/FLUX.1-schnell" and "black-forest-labs/FLUX.1-dev"
  • Tiling strategies can affect the processing time and quality of the transformed image

Common Issues

  • Missing API Key: Ensure the Hugging Face API token is set as HUGGINGFACE_HUB_ACCESS_TOKEN; instructions for that are in this guide
  • Memory Constraints: Large tile sizes or high inference steps may require significant memory resources