ComfyUI Node: OCS SimpleRestartSchedule

OCS SimpleRestartSchedule:

The OCS SimpleRestartSchedule node is designed to facilitate the creation of a Restart sampling schedule based on a text definition. This node is part of the Overly Complicated Sampling (OCS) suite and is specifically tailored to generate a Restart sampling schedule that can be used with OCS Sampler or RestartSampler nodes. The primary benefit of this node is its ability to interpret a text-based schedule definition, allowing for flexible and customizable sampling schedules. This can be particularly useful for AI artists looking to experiment with different sampling strategies to achieve unique artistic effects. The node ensures that the schedule is correctly parsed and applied, providing a robust and user-friendly way to manage complex sampling routines.

OCS SimpleRestartSchedule Input Parameters:

sigmas

The sigmas parameter represents the noise levels used in the sampling process. It is a crucial input as it defines the sequence of noise values that the sampling algorithm will use. The sigmas should be a one-dimensional tensor, and its length directly impacts the sampling process. If the dimensions are incorrect, the node will raise an error. This parameter does not have a default value and must be provided by the user.

start_step

The start_step parameter specifies the initial step from which the schedule will begin. It is an integer value with a default of 0. This parameter allows you to control the starting point of the sampling schedule, which can be useful for resuming or modifying existing sampling processes. The minimum value is 0, and there is no explicit maximum value, but it should be within the range of the sigmas length.

schedule

The schedule parameter is a text-based definition of the sampling schedule. It should be provided in JSON or YAML list format. This parameter allows for highly customizable schedules by defining intervals and jumps within the sampling process. If the schedule is not provided or is empty, the node will default to an empty list. The schedule must be correctly formatted, as improper formatting will result in a ValueError.

OCS SimpleRestartSchedule Output Parameters:

SIGMAS

The SIGMAS output parameter represents the modified sequence of noise levels after applying the Restart sampling schedule. This output is crucial for the subsequent sampling process, as it dictates the noise levels that will be used. The SIGMAS can be connected to an OCS Sampler or RestartSampler node, which will utilize the modified noise levels to perform the sampling. This output ensures that the customized schedule is effectively integrated into the sampling workflow.

OCS SimpleRestartSchedule Usage Tips:

• Ensure that the sigmas parameter is a one-dimensional tensor to avoid dimension-related errors.
• Use the start_step parameter to control the starting point of your schedule, which can be useful for resuming interrupted processes.
• Carefully format the schedule parameter in JSON or YAML list format to avoid parsing errors and ensure the schedule is applied correctly.
• Connect the SIGMAS output to an OCS Sampler or RestartSampler node to utilize the customized sampling schedule effectively.

OCS SimpleRestartSchedule Common Errors and Solutions:

ValueError: Schedule must be a JSON or YAML list

• Explanation: This error occurs when the schedule parameter is not provided in the correct format.
• Solution: Ensure that the schedule parameter is a properly formatted JSON or YAML list.

ValueError: Bad number of dimensions for sigmas

• Explanation: This error occurs when the sigmas parameter is not a one-dimensional tensor.
• Solution: Verify that the sigmas parameter is a one-dimensional tensor before passing it to the node.

ValueError: Schedule jump index out of range

• Explanation: This error occurs when a jump index in the schedule parameter is out of the valid range.
• Solution: Check the schedule parameter to ensure that all jump indices are within the valid range of the schedule length.

RuntimeError: Hit max iteration count. Loop in schedule?

• Explanation: This error occurs when the maximum iteration count is exceeded, possibly due to a loop in the schedule.
• Solution: Review the schedule parameter to ensure there are no loops or excessive iterations that could cause this error. Adjust the schedule to prevent infinite loops.