ComfyUI Node: ScaledCFGGuider

ScaledCFGGuider:

The ScaledCFGGuider node is designed to enhance the sampling process in AI art generation by providing a more nuanced control over the conditioning guidance. This node allows you to blend two different conditioning inputs with a specified strength, offering a more flexible and dynamic approach to generating images. By adjusting the conditioning factors and the strength of their influence, you can achieve a more refined and targeted output, making it easier to guide the model towards the desired artistic outcome. The main goal of this node is to provide a sophisticated mechanism for conditioning guidance, which can significantly improve the quality and specificity of the generated images.

ScaledCFGGuider Input Parameters:

model

This parameter specifies the model to be used for the sampling process. It is a required input and ensures that the node has the necessary model to perform the conditioning guidance.

cond1

This parameter represents the first conditioning input. It is used to guide the model towards a specific direction based on the provided conditioning data. This input is crucial for defining one aspect of the desired output.

cond2

This parameter represents the second conditioning input. Similar to cond1, it helps guide the model but allows for an additional layer of conditioning, providing more complexity and control over the generated output.

negative

This parameter represents the negative conditioning input. It is used to guide the model away from certain features or aspects, helping to refine the output by avoiding unwanted characteristics.

cfg

This parameter is a float value that controls the overall strength of the conditioning guidance. The default value is 8.0, with a minimum of 0.0 and a maximum of 100.0. Adjusting this value changes how strongly the model adheres to the provided conditioning inputs.

cond2_alpha

This parameter is a float value that determines the blending strength of the second conditioning input (cond2). The default value is 1.0, with a range from -100.0 to 100.0. This allows for fine-tuning the influence of cond2 relative to cond1.

ScaledCFGGuider Output Parameters:

GUIDER

The output of this node is a GUIDER object. This object encapsulates the configured guidance mechanism, which can then be used in the sampling process to generate images that adhere to the specified conditioning inputs and strengths. The GUIDER ensures that the model follows the defined guidance parameters, resulting in more controlled and targeted image generation.

ScaledCFGGuider Usage Tips:

• Experiment with different values for cfg to find the optimal strength for your specific use case. Higher values will make the model adhere more strictly to the conditioning inputs.
• Use cond2_alpha to balance the influence between cond1 and cond2. A positive value increases the influence of cond2, while a negative value decreases it.
• Combine cond1 and cond2 with complementary conditioning data to achieve more complex and nuanced outputs.

ScaledCFGGuider Common Errors and Solutions:

"Model not specified"

• Explanation: The model parameter is missing or not correctly specified.
• Solution: Ensure that you provide a valid model as input to the model parameter.

"Invalid conditioning input"

• Explanation: One or more of the conditioning inputs (cond1, cond2, negative) are not correctly specified.
• Solution: Verify that all conditioning inputs are correctly provided and are of the expected type.

"CFG value out of range"

• Explanation: The cfg parameter value is outside the allowed range (0.0 to 100.0).
• Solution: Adjust the cfg value to be within the specified range.

"cond2_alpha value out of range"

• Explanation: The cond2_alpha parameter value is outside the allowed range (-100.0 to 100.0).
• Solution: Adjust the cond2_alpha value to be within the specified range.