CustomControlNetWeights Node Documentation

1. Overview

The CustomControlNetWeights node is part of the ComfyUI-Advanced-ControlNet set of nodes. This node provides an advanced mechanism to manage the weights in a ControlNet. It facilitates the application of custom weights, allowing users to have granular control over the influence of ControlNet versus the main prompt. This is particularly useful for scenarios where different sections of a controlled image generation need varying levels of influence.

2. Inputs

The CustomControlNetWeights node accepts the following inputs:

• Base Multiplier (base_multiplier): This input allows you to adjust the softness of the weight applied to the ControlNet. It determines how much emphasis is placed on the user-specified custom weights versus the default settings.
• Unconditional Multiplier (uncond_multiplier): Controls the importance of the ControlNet compared to the original prompt. This can be seen as the inverse of the "My prompt is more important" feature, enabling users to set ControlNet effects to be more prominent. A value of 0.0 replicates the Auto1111 feature of making the ControlNet less influential, while values between 0.0 and 1.0 allow for more nuanced control.

3. Outputs

The CustomControlNetWeights node generates tailored weights based on the input parameters. The resulting output is a set of weights customized to influence the ControlNet's impact during image generation, making the output adaptable to the user's required emphasis.

4. Use in ComfyUI Workflows

In ComfyUI workflows, CustomControlNetWeights can be integrated as follows:

• Dynamic Image Styling: This node is useful in workflows where dynamic control over image style and aspect is needed. By adjusting weights, users can achieve varied results wherein certain parts of an image are influenced more heavily by the ControlNet while others follow the main prompt's guidance.
• Custom Effects: For advanced users wanting to replicate or closely achieve specific stylistic results, this node can be integrated to influence how closely the ControlNet follows specific parameters or influences particular aspects of the image.
• Integration with Keyframing: When used with Timestep and Latent Keyframes, this node provides fine control over image animation and transformation processes by altering weights as different keyframes are applied during sampling.

5. Special Features or Considerations

• Granular Control: One of the standout features is the granular control it provides over how much the ControlNet affects the output image compared to the primary image prompt. This allows for creative flexibility and precise adjustments for artists and developers.
• Compatibility: The node easily integrates with the rest of the Advanced-ControlNet workflow, including other nodes that allow for timestep and latent strength scheduling. This compatibility amplifies its utility across varied use cases.
• Auto1111 Feature Replication: The feature replicating the "My prompt is more important" and "ControlNet is more important" settings from Auto1111's sd-webui ControlNet extension provides familiarity for users who are accustomed to those dynamics.

When using this node, consideration should be given to the interplay between the base_multiplier and uncond_multiplier, as this balance determines the final visual weighting between the ControlNet and the prompt influence. Experimentation and adjustment are encouraged to achieve the desired outcome in creative projects.