Visit ComfyUI Online for ready-to-use ComfyUI environment
Enhances image sampling with tiled division for improved quality and performance.
The TiledKSamplerProvider
is a specialized node designed to enhance the sampling process by dividing the image into smaller tiles. This approach allows for more efficient and detailed sampling, especially useful for high-resolution images. By breaking down the image into manageable sections, the node ensures that each tile is processed with optimal settings, leading to improved overall quality and performance. This method is particularly beneficial for AI artists looking to generate high-quality images without overwhelming their system's resources. The TiledKSamplerProvider
leverages advanced sampling techniques and customizable parameters to provide a flexible and powerful tool for image generation.
The seed
parameter is an integer that initializes the random number generator used in the sampling process. It ensures reproducibility of results, meaning that using the same seed will produce the same output. The value ranges from 0 to 0xffffffffffffffff, with a default of 0.
The steps
parameter defines the number of sampling steps to be performed. More steps generally lead to higher quality images but require more processing time. The value ranges from 1 to 10000, with a default of 20.
The cfg
(Classifier-Free Guidance) parameter is a float that controls the strength of the guidance during sampling. Higher values result in images that more closely follow the provided conditioning. The value ranges from 0.0 to 100.0, with a default of 8.0.
The sampler_name
parameter specifies the sampling algorithm to be used. It is selected from a predefined list of samplers provided by comfy.samplers.KSampler.SAMPLERS
.
The scheduler
parameter determines the scheduling strategy for the sampling process. It is selected from a predefined list of schedulers provided by comfy.samplers.KSampler.SCHEDULERS
.
The denoise
parameter is a float that controls the amount of denoising applied during the sampling process. Higher values result in smoother images. The value ranges from 0.0 to 1.0, with a default of 1.0 and a step of 0.01.
The tile_width
parameter specifies the width of each tile in pixels. It allows for customization of the tile size to balance between detail and performance. The value ranges from 320 to MAX_RESOLUTION
, with a default of 512 and a step of 64.
The tile_height
parameter specifies the height of each tile in pixels. Similar to tile_width
, it allows for customization of the tile size. The value ranges from 320 to MAX_RESOLUTION
, with a default of 512 and a step of 64.
The tiling_strategy
parameter defines the method used to tile the image. Options include "random", "padded", and "simple", each offering different approaches to how the image is divided and processed.
The basic_pipe
parameter is a composite input that includes the model and conditioning information required for the sampling process. It typically contains the model, positive and negative conditioning, and other relevant data.
The KSAMPLER
output is the result of the tiled sampling process. It is an object that encapsulates the sampled image data, ready for further processing or final output. This output ensures that the image has been processed with the specified tiling and sampling parameters, providing a high-quality result.
tiling_strategy
options to find the best approach for your specific image and desired outcome.steps
parameter to balance between image quality and processing time, especially for high-resolution images.seed
parameter to reproduce specific results, which is useful for iterative design processes.'BNK_TiledKSampler' node isn't installed.
BNK_TiledKSampler
node is not installed in your environment.Tiled sampling for ComfyUI
extension from the provided GitHub repository link: https://github.com/BlenderNeko/ComfyUI_TiledKSampler
.Invalid value for parameter 'tile_width' or 'tile_height'.
tile_width
and tile_height
are within the range of 320 to MAX_RESOLUTION
and are multiples of 64.Unsupported sampler_name or scheduler.
comfy.samplers.KSampler.SAMPLERS
and comfy.samplers.KSampler.SCHEDULERS
.© Copyright 2024 RunComfy. All Rights Reserved.