⚓ Capitan Presets:
The CapitanShiftPresets node is designed to provide a convenient way to apply predefined configurations for generating sigma values, which are crucial in the sampling process of AI models. This node offers a selection of presets that cater to different needs, such as speed, balance, and quality, allowing you to easily choose the most suitable configuration for your specific task. By using these presets, you can streamline the process of setting up your model's sampling parameters, ensuring that you achieve the desired results efficiently. The node calculates sigma values based on the chosen preset, which influences the model's behavior during the sampling process, ultimately affecting the quality and characteristics of the generated output. This functionality is particularly beneficial for users who want to optimize their workflow without delving into the complexities of manually configuring each parameter.
⚓ Capitan Presets Input Parameters:
preset
The preset parameter allows you to select from a list of predefined configurations that determine the behavior of the sigma generation process. Each preset is tailored to achieve specific goals, such as speed, balance, or quality, by adjusting the number of steps, shift value, and configuration (cfg) settings. The available options include "Lumina2 (Fast)", "Lumina2 (Balanced)", "Lumina2 (Quality)", "Z-Image-Turbo (Speed)", "Z-Image-Turbo (Quality)", "Flux (Fast)", "Flux (Balanced)", "Flux (Quality)", "SD3 (Balanced)", and "SD3 (Quality)". By choosing a preset, you can quickly apply a set of parameters that are optimized for different scenarios, making it easier to achieve the desired results without manually adjusting each setting.
⚓ Capitan Presets Output Parameters:
sigmas
The sigmas output provides a sequence of sigma values that are used during the sampling process. These values are crucial for controlling the noise level and the progression of the sampling, directly impacting the quality and characteristics of the generated output. The sigma values are calculated based on the selected preset, ensuring that they align with the intended goals of speed, balance, or quality.
recommended_steps
The recommended_steps output indicates the number of steps that are suggested for the sampling process when using the selected preset. This value helps guide you in setting up the sampling process to align with the preset's intended performance characteristics, ensuring that the model operates efficiently and effectively.
recommended_cfg
The recommended_cfg output provides the recommended configuration (cfg) value for the sampling process. This value is part of the preset's configuration and plays a role in determining the behavior of the model during sampling. By following the recommended cfg, you can ensure that the model's settings are optimized for the chosen preset, contributing to the overall quality and efficiency of the output.
⚓ Capitan Presets Usage Tips:
- Choose a preset that aligns with your specific needs, such as "Lumina2 (Quality)" for high-quality outputs or "Z-Image-Turbo (Speed)" for faster results.
- Use the
recommended_stepsandrecommended_cfgoutputs to configure your model's sampling process, ensuring that you achieve optimal performance based on the selected preset.
⚓ Capitan Presets Common Errors and Solutions:
KeyError: '<preset_name>'
- Explanation: This error occurs when the specified preset name is not found in the predefined list of presets.
- Solution: Ensure that you select a preset from the available options: "Lumina2 (Fast)", "Lumina2 (Balanced)", "Lumina2 (Quality)", "Z-Image-Turbo (Speed)", "Z-Image-Turbo (Quality)", "Flux (Fast)", "Flux (Balanced)", "Flux (Quality)", "SD3 (Balanced)", or "SD3 (Quality)".
TypeError: 'NoneType' object is not subscriptable
- Explanation: This error may occur if the preset selection process fails, resulting in a
Nonevalue being used where a dictionary is expected. - Solution: Double-check that the preset parameter is correctly set and that the node is properly initialized with the available presets.