ComfyUI Node: Pre CFG perp-neg

Pre CFG perp-neg:

The Pre CFG perp-neg node is designed to enhance the conditioning process in AI models by adjusting the noise predictions based on perpendicular negative guidance. This node aims to refine the model's output by subtracting the influence of negative conditioning from the positive conditioning, thereby improving the overall quality and accuracy of the generated results. By leveraging the perpendicular component of the negative noise prediction, this node helps in achieving a more balanced and precise output, making it a valuable tool for AI artists looking to fine-tune their models for better performance.

Pre CFG perp-neg Input Parameters:

model

This parameter represents the AI model that will be used for the conditioning process. It is essential for the node to have access to the model to perform the necessary calculations and adjustments.

empty_conditioning

This parameter provides the conditioning data that is considered neutral or empty. It serves as a baseline for the node to compare and adjust the positive and negative conditioning.

neg_scale

This parameter controls the scaling factor for the negative conditioning. It determines the extent to which the negative conditioning will influence the final output. The value can range from 0.0 to 100.0, with a default value of 1.0. Adjusting this parameter allows you to fine-tune the impact of negative conditioning on the model's output.

Pre CFG perp-neg Output Parameters:

MODEL

The output of this node is the modified AI model with the adjusted conditioning. This model has been fine-tuned using the perpendicular negative guidance, resulting in improved performance and more accurate outputs.

Pre CFG perp-neg Usage Tips:

• Experiment with different neg_scale values to find the optimal balance between positive and negative conditioning for your specific use case.
• Use the empty_conditioning parameter to provide a neutral baseline that can help in achieving more precise adjustments.
• Regularly test the output model to ensure that the adjustments are leading to the desired improvements in performance and accuracy.

Pre CFG perp-neg Common Errors and Solutions:

"Model not found"

• Explanation: This error occurs when the specified model is not available or cannot be accessed by the node.
• Solution: Ensure that the model is correctly loaded and accessible before running the node.

"Invalid neg_scale value"

• Explanation: This error occurs when the neg_scale value is outside the acceptable range.
• Solution: Verify that the neg_scale value is within the range of 0.0 to 100.0 and adjust it accordingly.

"Empty conditioning data missing"

• Explanation: This error occurs when the empty_conditioning parameter is not provided or is invalid.
• Solution: Ensure that the empty_conditioning parameter is correctly set with valid conditioning data.