SoftControlNetWeights Node Documentation

Overview

The SoftControlNetWeights node is part of the ComfyUI-Advanced-ControlNet project. It allows users to apply custom weights to ControlNets in ComfyUI workflows, facilitating advanced control over the balancing of prompt importance and ControlNet effects. The node provides a mechanism to replicate features from the Auto1111 sd-webui ControlNet extension, specifically the "My prompt is more important" and "ControlNet is more important" functionalities. By adjusting the base_multiplier and uncond_multiplier, users can fine-tune how these effects manifest across different sampling scenarios.

Functionality

1. Replicate Auto1111 Features: Implements functionalities similar to the Auto1111 extension, giving granular control over the balance between prompt importance and ControlNet influence.
2. Custom Weight Application: Allows the application of custom weights which can be varied across different steps of the sampling process.
3. Support for Various Models: Compatible with ControlNets, T2IAdapters, ControlLoRAs, and other related models, providing flexibility in applying soft weights.

Inputs

• base_multiplier (🟦 Optional Widget/Input): Adjusts the softness of the weights. Higher values increase prompt importance, while lower values enhance ControlNet influence.
• uncond_multiplier (🟦 Optional Widget/Input): Controls the granular application of the "ControlNet is more important" feature. A value of 0.0 replicates the identical results from the Auto1111 implementation, but other values within the range of 0.0 to 1.0 can be used for more nuanced control.
• control_net (🟩 Required Input): The loaded ControlNet for which these weights are intended. This input is essential for converting and applying the weights appropriately.

Outputs

• weighted_control_net (🟪 Output): The resultant ControlNet with applied soft weights. This can be used in subsequent nodes to render the desired influence as determined by the input parameters.

Usage in ComfyUI Workflows

The SoftControlNetWeights node is typically integrated into workflows where dynamic adjustments to the influence of ControlNets are necessary. Here's how it might be used:

1. Load a ControlNet: Start by loading a ControlNet using a compatible node (like Load Advanced ControlNet Model).
2. Configure Weights: Using the SoftControlNetWeights node, configure the base_multiplier and uncond_multiplier to set the desired balance of prompt importance.
3. Apply to ControlNet: The output of the SoftControlNetWeights node— the weighted ControlNet—can then be fed into other advanced ControlNet application nodes.
4. Sequential Processing: This weighted ControlNet can be scheduled across sampling steps and latents using other nodes like Timestep Keyframe nodes for even more control.

Special Features or Considerations

• Granular Control: Offers highly detailed control over the influence balances in generated outputs by manipulating the base and uncond multipliers.
• Versatility with Models: Works seamlessly with a range of models beyond standard ControlNets, making it suitable for complex workflows in ComfyUI.
• Flexible Integration: Can be used alongside timestep keyframe scheduling to alternate influences throughout the sampling process, empowering users to achieve highly customized outputs.

As part of a larger suite of advanced nodes, SoftControlNetWeights excels in scenarios requiring fine-tuning and sophisticated control over how visual outputs incorporate and balance different influences.