ComfyUI Node: Pre CFG channel multiplier

Pre CFG channel multiplier:

The Pre CFG channel multiplier node is designed to adjust the influence of specific channels in your model's conditioning process. This node allows you to apply different scaling factors to individual channels, providing fine-grained control over how each channel contributes to the final output. By selectively amplifying or diminishing the impact of certain channels, you can achieve more nuanced and targeted modifications to your model's behavior. This is particularly useful in scenarios where certain features or attributes need to be emphasized or suppressed, enhancing the overall quality and specificity of the generated content.

Pre CFG channel multiplier Input Parameters:

model

This parameter represents the model you are working with. It is essential for the node to know which model to apply the channel multipliers to.

channel_1

This parameter specifies the scaling factor for the first channel. The value you set here will multiply the output of the first channel, allowing you to control its influence. The default value is typically 1.0, meaning no change.

channel_2

This parameter specifies the scaling factor for the second channel. Similar to channel_1, it allows you to adjust the influence of the second channel by multiplying its output. The default value is 1.0.

channel_3

This parameter specifies the scaling factor for the third channel. By setting this value, you can control the contribution of the third channel to the final output. The default value is 1.0.

channel_4

This parameter specifies the scaling factor for the fourth channel. Adjusting this value will multiply the output of the fourth channel, allowing you to fine-tune its impact. The default value is 1.0.

selection

This parameter determines which channels to apply the multipliers to. The options are "both", "cond", and "uncond". "both" applies the multipliers to both conditional and unconditional channels, "cond" applies only to conditional channels, and "uncond" applies only to unconditional channels.

start_at_sigma

This parameter sets the starting sigma value at which the channel multipliers will begin to take effect. It allows you to control the range of sigma values over which the multipliers are applied. The default value is typically set to a high value like 1000.0.

end_at_sigma

This parameter sets the ending sigma value at which the channel multipliers will stop being applied. It defines the lower bound of the sigma range for the multipliers. The default value is usually set to a lower value like 1.0.

Pre CFG channel multiplier Output Parameters:

MODEL

The output is the modified model with the applied channel multipliers. This model will now use the specified scaling factors for the selected channels during the conditioning process, allowing for more controlled and targeted modifications to the generated content.

Pre CFG channel multiplier Usage Tips:

• To emphasize specific features in your generated content, increase the scaling factor for the corresponding channels.
• Use the "selection" parameter to target only the conditional or unconditional channels if you want to apply changes selectively.
• Adjust the start_at_sigma and end_at_sigma parameters to control the range of sigma values over which the multipliers are applied, ensuring that the changes take effect only within the desired range.

Pre CFG channel multiplier Common Errors and Solutions:

WRONG CHANNEL SELECTED. THE LATENT SPACE ONLY HAS {len(conds_out[1][b])} CHANNELS

• Explanation: This error occurs when you try to apply a multiplier to a channel that does not exist in the latent space of the model.
• Solution: Ensure that the channel number you specify is within the range of available channels in the model's latent space. Adjust the channel number to a valid value.