ComfyUI Node: BlehRefinerAfter

BlehRefinerAfter:

The BlehRefinerAfter node is designed to refine and enhance the performance of AI models by dynamically switching between a primary model and a refiner model based on a specified time mode and start time. This node is particularly useful for AI artists who want to improve the quality of their generated outputs by leveraging different models at different stages of the generation process. The main goal of this node is to provide a seamless transition between models, ensuring that the most appropriate model is used at each stage to achieve the best possible results.

BlehRefinerAfter Input Parameters:

time_mode

The time_mode parameter determines the method used to evaluate the transition point between the primary model and the refiner model. It accepts three options: "timestep", "percent", and "sigma". Each option represents a different way to measure time during the model's execution. "timestep" uses discrete time steps, "percent" uses a percentage of the total process, and "sigma" uses the sigma value. Choosing the appropriate time mode is crucial for ensuring the correct model is applied at the right time.

start_time

The start_time parameter specifies the point at which the transition between the primary model and the refiner model occurs. The value of start_time depends on the selected time_mode. For "timestep", it ranges from 0.0 to 999.0 with a default of 199.0. For "percent", it ranges from 0.0 to 1.0. For "sigma", it is based on the sigma values of the model. Setting this parameter correctly ensures that the transition happens at the optimal point for model refinement.

model

The model parameter is the primary AI model used for the initial stages of the generation process. This model is applied until the specified start_time is reached, after which the refiner model may take over. The primary model should be chosen based on its performance in the early stages of the generation process.

refiner_model

The refiner_model parameter is the secondary AI model used to refine and enhance the output after the specified start_time is reached. This model is designed to improve the quality of the generated output by providing additional refinement and detail. The refiner model should be selected based on its ability to enhance the output in the later stages of the generation process.

BlehRefinerAfter Output Parameters:

MODEL

The output parameter MODEL represents the AI model that has been dynamically switched between the primary model and the refiner model based on the specified time_mode and start_time. This output ensures that the most appropriate model is used at each stage of the generation process, resulting in higher quality and more refined outputs.

BlehRefinerAfter Usage Tips:

• To achieve the best results, carefully choose the time_mode and start_time parameters based on the characteristics of your primary and refiner models.
• Experiment with different start_time values to find the optimal transition point that maximizes the quality of your generated outputs.
• Ensure that both the primary model and the refiner model are well-suited for their respective stages of the generation process to achieve seamless transitions and high-quality results.

BlehRefinerAfter Common Errors and Solutions:

BlehRefinerAfter: invalid value for percent start time

• Explanation: This error occurs when the start_time value for the "percent" time_mode is set outside the valid range of 0.0 to 1.0.
• Solution: Ensure that the start_time value is within the range of 0.0 to 1.0 when using the "percent" time_mode.

BlehRefinerAfter: invalid time mode

• Explanation: This error occurs when an invalid time_mode is specified.
• Solution: Ensure that the time_mode parameter is set to one of the valid options: "timestep", "percent", or "sigma".

Model loading issues

• Explanation: This error occurs when the primary or refiner model fails to load correctly.
• Solution: Verify that the models are correctly specified and available. Ensure that the models are compatible with the node and are properly loaded into the system.