Diversity Boost (V3):
DiversityBoostCoreV3 is a sophisticated node designed to enhance the composition diversity of distilled models, particularly in the realm of AI-generated art. This node operates by implementing a combination of high-frequency (HF) attenuation and discrete cosine transform (DCT) composition push, specifically at the initial step of the process. The primary goal of DiversityBoostCoreV3 is to restore and maintain the richness and variety in compositions that might otherwise be lost during the distillation process of models. By doing so, it ensures that the generated outputs are not only diverse but also maintain a high level of detail and complexity. This node is particularly beneficial for artists and creators who seek to push the boundaries of creativity and achieve unique artistic expressions through AI.
Diversity Boost (V3) Input Parameters:
strength
The strength parameter controls the intensity of the diversity boost applied to the model. A higher value increases the effect, potentially leading to more varied and complex compositions. The default value is 2.0, with a typical range from 0.0 to 5.0, allowing you to fine-tune the diversity level according to your artistic needs.
clamp_val
clamp_val determines the maximum allowable value for the diversity boost effect, effectively capping the intensity to prevent over-saturation or distortion. The default is set at 0.5, with a range from 0.0 to 1.0, ensuring that the output remains within a visually appealing spectrum.
noise_type
This parameter specifies the type of noise to be used in the diversity boosting process. The default option is "pink," which is known for its balanced frequency distribution, but other types may be available depending on the implementation, each offering different aesthetic effects.
dc_preserve
dc_preserve is used to maintain the direct current (DC) component of the image, which can help in preserving the overall brightness and contrast. The default value is 0.0, with a range from 0.0 to 1.0, allowing you to adjust how much of the original image's tonal balance is retained.
energy_compensate
This boolean parameter, when enabled, compensates for energy loss during the diversity boosting process, ensuring that the output maintains a consistent level of brightness and contrast. It is set to False by default.
mode
The mode parameter defines the method used for diversity boosting. The default is "polynomial," which provides a smooth and continuous effect, but other modes like "butterworth" may be available for different artistic outcomes.
hf_factor
hf_factor adjusts the influence of high-frequency components in the diversity boosting process. A higher value increases the prominence of fine details. The default is 1.0, with a range from 0.0 to 2.0.
lf_factor
This parameter controls the impact of low-frequency components, affecting the overall structure and form of the composition. The default value is 0.3, with a range from 0.0 to 1.0.
transition
transition defines the smoothness of the transition between different frequency components. The default is 2.0, with a range from 0.0 to 5.0, allowing for either sharp or gradual transitions.
schedule
The schedule parameter determines the progression of the diversity boost effect over time. The default is "linear," providing a steady increase, but other schedules may be available for varied temporal dynamics.
Diversity Boost (V3) Output Parameters:
modified
The modified output represents the final image or video after the diversity boost has been applied. This output is crucial as it embodies the enhanced diversity and complexity achieved through the node's processing, offering a richer and more varied artistic expression.
Diversity Boost (V3) Usage Tips:
- Experiment with the
strengthparameter to find the right balance between diversity and coherence in your compositions. - Use
clamp_valto prevent over-exaggeration of effects, ensuring that the output remains visually appealing. - Try different
noise_typesettings to explore various artistic styles and textures. - Adjust
hf_factorandlf_factorto emphasize either fine details or overall structure, depending on your artistic goals.
Diversity Boost (V3) Common Errors and Solutions:
"ValueError: Invalid noise type"
- Explanation: This error occurs when an unsupported noise type is specified.
- Solution: Ensure that the
noise_typeparameter is set to a valid option, such as "pink."
"RuntimeError: Clamp value out of range"
- Explanation: The
clamp_valparameter is set outside the acceptable range. - Solution: Adjust
clamp_valto be within the range of 0.0 to 1.0.
"TypeError: Unsupported mode"
- Explanation: The
modeparameter is set to an unsupported value. - Solution: Verify that the
modeis set to a supported option like "polynomial" or "butterworth."
