ComfyUI Node: Runtime44 Tiled Mask Sampler

Runtime44 Tiled Mask Sampler:

The Runtime44TiledMaskSampler is a specialized node designed to facilitate the use of masks during a tiled sampling process. This node is particularly useful for AI artists who need to apply specific masks to different sections of an image or latent space, ensuring that the sampling process respects these masks. By breaking down the image into tiles and applying the mask to each tile, this node allows for more precise control over the sampling process, leading to higher quality and more targeted results. The primary goal of this node is to enhance the flexibility and accuracy of the sampling process, making it easier to achieve the desired artistic effects.

Runtime44 Tiled Mask Sampler Input Parameters:

model

This parameter specifies the model to be used for the sampling process. It is essential as it defines the underlying architecture and weights that will generate the output. The model parameter ensures that the sampling process is consistent with the chosen model's capabilities and characteristics.

positive

This parameter represents the positive conditioning input, which guides the model towards generating desired features in the output. It is crucial for steering the sampling process in a direction that aligns with the artist's vision.

negative

This parameter represents the negative conditioning input, which helps the model avoid generating unwanted features. By providing negative examples, the sampling process can be fine-tuned to exclude specific characteristics from the output.

latent

The latent parameter is the initial latent space representation that will be modified during the sampling process. It serves as the starting point for the generation and is essential for defining the initial conditions of the sampling.

mask

The mask parameter is a tensor that defines the areas of the image or latent space that should be influenced during the sampling process. It allows for selective modification, ensuring that only specific regions are altered according to the mask's design.

seed

This integer parameter sets the random seed for the sampling process, ensuring reproducibility. The default value is 0, with a minimum of 0 and a maximum of 0xFFFFFFFFFFFFFFFF. Setting the seed allows for consistent results across different runs.

steps

This integer parameter defines the number of steps to be taken during the sampling process. The default value is 20, with a minimum of 1 and a maximum of 10000. More steps generally lead to higher quality results but require more computational resources.

cfg

The cfg parameter is a float that controls the classifier-free guidance scale. The default value is 8.0, with a range from 0.0 to 100.0 and a step size of 0.1. This parameter balances the trade-off between adhering to the conditioning inputs and exploring new variations.

sampler_name

This parameter specifies the name of the sampler to be used. It determines the sampling algorithm, which can significantly impact the quality and style of the generated output.

scheduler

The scheduler parameter defines the scheduling strategy for the sampling process. Different schedulers can affect the convergence and quality of the results.

denoise

This float parameter controls the denoising strength during the sampling process. The default value is 1.0, with a range from 0.0 to 1.0 and a step size of 0.01. Adjusting this parameter can help in reducing noise and improving the clarity of the output.

mask_feather

This integer parameter specifies the feathering amount for the mask, which smooths the edges of the mask. The default value is 13, with a range from 0 to 10000. Feathering helps in blending the masked regions more naturally with the rest of the image.

mask_dilation

This integer parameter controls the dilation of the mask, which expands or contracts the masked areas. The default value is 0, with a range from -10000 to 10000. Dilation can be used to adjust the influence area of the mask.

Runtime44 Tiled Mask Sampler Output Parameters:

LATENT

The output parameter is a latent space representation that has been modified according to the specified mask and sampling process. This latent output can be further processed or directly used to generate the final image. It encapsulates the changes made during the tiled sampling process, reflecting the applied masks and conditioning inputs.

Runtime44 Tiled Mask Sampler Usage Tips:

• To achieve the best results, carefully design your mask to target specific areas of the image that need modification. This will ensure that the sampling process focuses on the desired regions.
• Experiment with different values for the steps and cfg parameters to find the optimal balance between quality and computational efficiency. More steps generally lead to better results but require more processing time.
• Use the mask_feather parameter to smooth the edges of your mask, which can help in creating more natural transitions between masked and unmasked areas.

Runtime44 Tiled Mask Sampler Common Errors and Solutions:

Mask tensor shape mismatch

• Explanation: This error occurs when the shape of the mask tensor does not match the expected dimensions.
• Solution: Ensure that the mask tensor has the correct shape and dimensions as required by the node. Adjust the mask tensor to match the expected input size.

Invalid seed value

• Explanation: This error occurs when the seed value is outside the acceptable range.
• Solution: Verify that the seed value is within the range of 0 to 0xFFFFFFFFFFFFFFFF. Adjust the seed value to fall within this range.

Insufficient steps

• Explanation: This error occurs when the number of steps is set too low, leading to poor quality results.
• Solution: Increase the number of steps to improve the quality of the output. A higher number of steps generally leads to better results.

Mask dilation out of range

• Explanation: This error occurs when the mask dilation value is set outside the acceptable range.
• Solution: Ensure that the mask dilation value is within the range of -10000 to 10000. Adjust the dilation value accordingly.