ComfyUI Node: SamplerSupreme

SamplerSupreme:

SamplerSupreme is a sophisticated node designed to enhance the sampling process in AI art generation. It leverages advanced techniques to provide high-quality, detailed outputs by incorporating various methods such as noise modulation, edge enhancement, and normalization. This node is particularly beneficial for artists looking to achieve more refined and controlled results in their generative art projects. By utilizing a combination of step and substep methods, along with customizable parameters, SamplerSupreme offers a versatile and powerful tool for fine-tuning the sampling process, ensuring that the generated images meet the desired artistic standards.

SamplerSupreme Input Parameters:

model

This parameter specifies the model to be used for sampling. It is essential for defining the architecture and weights that will guide the sampling process.

x

This parameter represents the initial input tensor for the sampling process. It serves as the starting point for generating the output image.

sigmas

Sigmas are used to control the noise levels during the sampling process. They play a crucial role in determining the amount of detail and texture in the generated image.

extra_args

This optional parameter allows you to pass additional arguments to the sampling function, providing further customization and control over the process.

callback

An optional parameter that can be used to specify a callback function, which will be called at each step of the sampling process. This can be useful for monitoring progress or making real-time adjustments.

disable

This parameter can be used to disable certain features or steps in the sampling process, offering a way to streamline or simplify the operation.

s_noise

Controls the strength of the noise applied during sampling. The default value is 1.0, and it can be adjusted to achieve different levels of noise intensity.

noise_sampler_type

Specifies the type of noise sampler to be used. The default is "gaussian," but other types can be selected based on the desired effect.

noise_sampler

An optional parameter that allows you to provide a custom noise sampler. If not specified, a default noise sampler will be used.

eta

Controls the amount of noise added at each step. The default value is 1.0, and it can be adjusted to fine-tune the noise application.

step_method

Defines the method to be used for each sampling step. The default is "euler," but other methods can be selected to achieve different results.

substep_method

Specifies the method to be used for substeps within each sampling step. The default is "euler," but other methods can be chosen for more control.

centralization

Controls the centralization of the noise. The default value is 0.05, and it can be adjusted to modify the noise distribution.

normalization

Adjusts the normalization of the noise. The default value is 0.05, and it can be fine-tuned to achieve the desired effect.

edge_enhancement

Enhances the edges in the generated image. The default value is 0.25, and it can be adjusted to emphasize or soften edges.

perphist

Controls the peripheral histogram equalization. The default value is 0.5, and it can be adjusted to balance the histogram distribution.

substeps

Specifies the number of substeps to be used in each sampling step. The default value is 2, and it can be increased for more detailed sampling.

noise_modulation

Defines the type of noise modulation to be applied. The default is "intensity," but other types can be selected based on the desired effect.

modulation_strength

Controls the strength of the noise modulation. The default value is 2.0, and it can be adjusted to achieve different modulation effects.

modulation_dims

Specifies the number of dimensions for the noise modulation. The default value is 3, and it can be adjusted to control the complexity of the modulation.

reversible_eta

Controls the reversibility of the eta parameter. The default value is 1.0, and it can be adjusted to fine-tune the reversibility.

SamplerSupreme Output Parameters:

SAMPLER

The output of the SamplerSupreme node is a sampler object that encapsulates all the specified parameters and methods. This sampler can be used to generate high-quality, detailed images based on the input parameters and the chosen model.

SamplerSupreme Usage Tips:

• Experiment with different noise_sampler_type settings to achieve various artistic effects.
• Adjust the edge_enhancement parameter to emphasize or soften edges in your generated images.
• Use the substeps parameter to increase the level of detail in your samples, especially for complex scenes.
• Fine-tune the centralization and normalization parameters to control the noise distribution and achieve a balanced output.

SamplerSupreme Common Errors and Solutions:

"Invalid model specified"

• Explanation: The model parameter is not correctly specified or is missing.
• Solution: Ensure that you provide a valid model that is compatible with the SamplerSupreme node.

"Noise sampler type not recognized"

• Explanation: The noise_sampler_type parameter is set to an unrecognized value.
• Solution: Verify that the noise_sampler_type is set to a valid option, such as "gaussian."

"Callback function error"

• Explanation: The callback function provided in the callback parameter is causing an error.
• Solution: Check the callback function for any issues and ensure it is correctly implemented.

"Parameter out of range"

• Explanation: One or more parameters are set to values outside their acceptable ranges.
• Solution: Review the parameter values and ensure they fall within the specified minimum and maximum limits.