ComfyUI Node: Tiled Upscaler Script

Tiled Upscaler Script:

The Tiled Upscaler Script is designed to enhance the resolution of images by dividing them into smaller tiles and processing each tile individually. This method allows for efficient upscaling without compromising the quality of the image, especially when dealing with high-resolution outputs. The script supports various tiling strategies and can integrate with ControlNet for more refined control over the upscaling process. By using this node, you can achieve high-quality upscaling results while managing memory usage effectively, making it ideal for AI artists looking to enhance their artwork with minimal technical hassle.

Tiled Upscaler Script Input Parameters:

upscale_by

This parameter determines the factor by which the image will be upscaled. It controls the overall increase in resolution. The value must be between 1 and 2, with the default being 1.5. Adjusting this parameter will directly impact the final resolution of the upscaled image.

tile_size

Specifies the size of each tile in pixels. This parameter is crucial for managing memory usage and processing time. Smaller tile sizes may reduce memory consumption but increase processing time, while larger tiles may do the opposite. The default value is 256 pixels.

tiling_strategy

Defines the strategy used for tiling the image. Options include "none", "overlap", and "non-overlap". The chosen strategy affects how tiles are processed and stitched back together. For instance, "overlap" can help reduce visible seams between tiles.

tiling_steps

Indicates the number of steps to be taken during the tiling process. This parameter influences the granularity of the tiling operation. More steps can lead to finer control but may increase processing time. The default value is 10 steps.

seed

A numerical value used to initialize the random number generator for reproducibility. Using the same seed will produce the same upscaling results, which is useful for consistency in batch processing.

denoise

Controls the level of noise reduction applied during the upscaling process. A higher value results in smoother images but may lose some details. The default value is 0.5.

use_controlnet

A boolean parameter that determines whether ControlNet should be used for additional control over the upscaling process. If set to true, ControlNet will be applied, providing more refined results.

tile_controlnet

Specifies the ControlNet model to be used if use_controlnet is enabled. This parameter allows for the selection of different ControlNet models to suit various upscaling needs.

strength

Determines the strength of the upscaling effect. Higher values result in more pronounced upscaling, while lower values produce subtler enhancements. The default value is 0.8.

script

An optional dictionary parameter that can be used to pass additional settings or configurations to the upscaling process. This allows for further customization and fine-tuning of the results.

Tiled Upscaler Script Output Parameters:

script

The output is a dictionary containing the processed script with the applied upscaling settings. This output can be used for further processing or to review the applied configurations.

Tiled Upscaler Script Usage Tips:

• Experiment with different tile_size values to find the optimal balance between memory usage and processing time for your specific hardware.
• Use the seed parameter to ensure reproducibility when processing multiple images with the same settings.
• Enable use_controlnet for more refined control over the upscaling process, especially when working with complex images.
• Adjust the denoise parameter to find the right balance between noise reduction and detail preservation in your upscaled images.

Tiled Upscaler Script Common Errors and Solutions:

"HighRes-Fix Warning: When using 'ttl_nn.<latent_upscaler>', 'upscale_by' must be between 1 and 2. Rounding to the nearest valid value."

• Explanation: The upscale_by value provided is outside the acceptable range.
• Solution: Ensure that the upscale_by parameter is set between 1 and 2.

"HighRes-Fix Warning: When using 'city96.<latent_upscaler>', 'upscale_by' must be one of [valid values]. Rounding to the nearest valid value."

• Explanation: The upscale_by value provided is not one of the valid options for the city96 upscaler.
• Solution: Use one of the valid upscale_by values specified in the warning message.

"Chosen latent upscale method not found! defaulting to '<default_upscaler>'."

• Explanation: The specified latent upscaler method is not recognized.
• Solution: Ensure that the latent_upscaler parameter is set to a valid method or accept the default method provided in the warning.