Flip Sigmas Adjusted(FlipSigmasAdjusted)
This node technically first completely reverses the order of the input sigmas sequence.
Flip Sigmas Adjusted
The Flip Sigmas Adjusted node technically first completely reverses the order of the input sigmas sequence.
The Flip Sigmas Adjusted node is a utility node for processing and adjusting noise schedule (sigmas) sequences. Its core function is to receive a sigmas sequence, perform operations such as reversal, offset, and scaling on it, thereby generating a new, adjusted noise intensity sequence. It is commonly used to control the denoising steps in the image generation process.
Node Functionality
Technically, this node first completely reverses the order of the input sigmas sequence. Subsequently, it allows users to apply an index offset to the elements in the sequence via the offset_by parameter to alter the evolution path of the noise intensity. Users can also choose whether to normalize the entire sequence using its last value or perform global scaling via the divide_by parameter. All operations ensure that no zero values appear in the sequence to avoid computational errors.
Node Parameter Description - Flip Sigmas Adjusted
Connection Inputs (Inputs)
| Parameter Name | Data Type | Required | Default Value | Value Range/Options | Description |
|---|---|---|---|---|---|
sigmas | SIGMAS | Yes | - | - | sigmas (Input Parameter): The input original noise schedule sequence to be processed. This sequence is typically generated by a scheduler node and defines the intensity for each step of the denoising process. |
Control Parameters (Parameters)
| Parameter Name | Data Type | Required | Default Value | Value Range/Options | Description |
|---|---|---|---|---|---|
divide_by_last_sigma | BOOLEAN | Yes | False | - | divide_by_last_sigma (Input Parameter): A boolean switch that determines whether to normalize the entire sequence using the last sigma value of the sequence. When enabled, all values will be divided by the last value. |
divide_by | FLOAT | Yes | 1 | 1 - 255 (Step: 0.01) | divide_by (Input Parameter): A divisor used for globally scaling the sequence. All adjusted sigma values will be divided by this number. The value range is 1 to 255, with a step size of 0.01. |
offset_by | INT | Yes | 1 | -100 - 100 (Step: 1) | offset_by (Input Parameter): Sets the index offset for the sequence elements. A positive value causes the current element to take the value from an earlier position (smaller index). The value range is -100 to 100, with a step size of 1. |
Output
| Parameter Name | Data Type | Description |
|---|---|---|
| SIGMAS | SIGMAS | The new noise schedule sequence generated after operations such as offset and scaling. |
| sigmas_string | STRING | sigmas_string (Output Parameter): The new sequence values output as a string, facilitating direct viewing or logging by the user. |
Usage Scenarios
In image generation workflows, this node is often placed before a KSampler to modify the noise schedule used by the sampler. For example, by adjusting the offset_by value, one can create denoising trajectories with unique rhythms different from the standard reverse diffusion process, thereby influencing the style or detail generation process of the final image.
Notes
When using this node, note that the input sigmas sequence usually comes from other scheduler nodes (e.g., BasicScheduler). Excessively large offset_by values may cause a large number of elements at the beginning or end of the sequence to be set to very small values, potentially making the sampling process unstable.
Flip Sigmas Adjusted Node Source Code Link
The Flip Sigmas Adjusted node is from the ComfyUI-KJNodes node package.
Comments
Sign in with GitHub to join the discussion.