ComfyUI Node: Primere LCM selector

Primere LCM selector:

The PrimereLCMSelector node is designed to facilitate the selection between two different model concepts: the standard model and the Latent Concept Model (LCM). This node allows you to switch between these modes seamlessly, adjusting various parameters to optimize the performance and output quality based on the selected model concept. By toggling the use_lcm parameter, you can dynamically change the sampler, scheduler, steps, and configuration scale to suit the specific needs of your project. This flexibility is particularly beneficial for AI artists who want to experiment with different model behaviors without manually adjusting multiple settings each time.

Primere LCM selector Input Parameters:

use_lcm

This boolean parameter determines whether the Latent Concept Model (LCM) mode is activated. When set to True, the node switches to LCM mode, adjusting the sampler, scheduler, steps, and configuration scale accordingly. The default value is False.

sampler_name

This parameter specifies the name of the sampler to be used when use_lcm is set to False. It defines the algorithm for sampling during the model's execution. The default value is euler.

scheduler_name

This parameter sets the name of the scheduler to be used when use_lcm is set to False. The scheduler controls the sequence and timing of operations within the model. The default value is normal.

lcm_sampler_name

This parameter specifies the name of the sampler to be used when use_lcm is set to True. It defines the algorithm for sampling in LCM mode. The default value is lcm.

lcm_scheduler_name

This parameter sets the name of the scheduler to be used when use_lcm is set to True. The scheduler controls the sequence and timing of operations within the model in LCM mode. The default value is sgm_uniform.

cfg_scale

This parameter defines the configuration scale to be used when use_lcm is set to False. It influences the model's behavior and output quality. The default value is 7.

steps

This parameter specifies the number of steps to be used when use_lcm is set to False. It determines the number of iterations the model will perform. The default value is 12.

lcm_cfg_scale

This parameter defines the configuration scale to be used when use_lcm is set to True. It influences the model's behavior and output quality in LCM mode. The default value is 1.2.

lcm_steps

This parameter specifies the number of steps to be used when use_lcm is set to True. It determines the number of iterations the model will perform in LCM mode. The default value is 6.

Primere LCM selector Output Parameters:

sampler_name

This output parameter returns the name of the sampler that was selected based on the use_lcm setting. It helps in understanding which sampling algorithm was applied during the model's execution.

scheduler_name

This output parameter returns the name of the scheduler that was selected based on the use_lcm setting. It indicates the sequence and timing control used within the model.

steps

This output parameter returns the number of steps that were used based on the use_lcm setting. It shows the number of iterations the model performed.

cfg_scale

This output parameter returns the configuration scale that was used based on the use_lcm setting. It reflects the influence on the model's behavior and output quality.

model_concept

This output parameter returns the model concept that was selected, either Normal or LCM. It provides clarity on which model concept was applied during the execution.

Primere LCM selector Usage Tips:

• To quickly switch between standard and LCM modes, toggle the use_lcm parameter. This will automatically adjust the relevant settings for you.
• Experiment with different cfg_scale and steps values in both modes to find the optimal configuration for your specific project needs.

Primere LCM selector Common Errors and Solutions:

Invalid sampler or scheduler name

• Explanation: The provided sampler or scheduler name does not match any known algorithms.
• Solution: Ensure that the sampler_name, scheduler_name, lcm_sampler_name, and lcm_scheduler_name parameters are set to valid names recognized by the model.

Steps value out of range

• Explanation: The number of steps specified is either too low or too high for the model to process effectively.
• Solution: Adjust the steps and lcm_steps parameters to values within a reasonable range, typically between 1 and 100.

Configuration scale too high or too low

• Explanation: The cfg_scale or lcm_cfg_scale values are set to extremes that the model cannot handle.
• Solution: Set the cfg_scale and lcm_cfg_scale parameters to moderate values, usually between 0.1 and 10, to ensure stable model performance.