ComfyUI Node: Weight Lifter 🏋️‍♀️

Weight Lifter 🏋️‍♀️:

The WeightLifter node is designed to apply systematic or random weights to tags within a given context, offering a versatile tool for AI artists to manipulate and enhance their creative outputs. This node provides several modes of operation, including RANDOM, GRADIENT, BURST, and NOISE, each offering a unique method of weight application. RANDOM mode assigns independent random weights within a specified range, while GRADIENT mode ensures weights progress smoothly across tags. BURST mode focuses on clusters of tags, applying stronger weights near random centers, and NOISE mode introduces smooth low-pass noise to selected tags. Additionally, the node offers keyword handling options such as NONE, ONLY, IGNORE, BOOST, and SUPPRESS, allowing for precise control over how keywords are treated. Existing tag behavior can be managed through PRESERVE, MODIFY, or OVERWRITE settings, ensuring flexibility in how weights are applied to tags with pre-existing weights. The node maintains the exact formatting of whitespace and newlines, and defaults the baseline to the midpoint unless specified otherwise. This functionality allows for nuanced control over tag weighting, enabling artists to fine-tune their creative processes with ease.

Weight Lifter 🏋️‍♀️ Input Parameters:

mode

The mode parameter determines the method by which weights are applied to tags. Options include RANDOM, GRADIENT, BURST, and NOISE, each offering a distinct approach to weight distribution. RANDOM mode applies independent random weights within a specified range, GRADIENT mode ensures a smooth progression of weights, BURST mode targets clusters of tags, and NOISE mode introduces smooth low-pass noise. This parameter significantly impacts the distribution and effect of weights on the tags.

keyword_mode

The keyword_mode parameter specifies how keywords are handled during the weight application process. Options include NONE, ONLY, IGNORE, BOOST, and SUPPRESS. NONE applies no special rules to keywords, ONLY modifies keywords, IGNORE skips keywords, BOOST forces keywords near the maximum weight, and SUPPRESS forces non-keywords near the minimum weight. This parameter allows for targeted manipulation of keywords, enhancing or diminishing their influence as desired.

existing_tag_behavior

The existing_tag_behavior parameter controls how tags with pre-existing weights are treated. Options include PRESERVE, MODIFY, and OVERWRITE. PRESERVE ensures existing weights remain unchanged, MODIFY adjusts the existing weight by a calculated delta, and OVERWRITE replaces the existing weight with the proposed weight. This parameter provides flexibility in managing tags with pre-existing weights, allowing for either preservation or modification.

min_weight

The min_weight parameter sets the minimum weight that can be applied to a tag. This value ensures that weights do not fall below a specified threshold, maintaining a baseline level of influence for all tags. The minimum weight is crucial for preventing tags from being overly diminished in the weighting process.

max_weight

The max_weight parameter defines the maximum weight that can be applied to a tag. This value caps the influence a tag can have, preventing any single tag from becoming disproportionately weighted. The maximum weight is essential for maintaining balance and preventing excessive emphasis on certain tags.

keyword_variance

The keyword_variance parameter introduces variability in the weight applied to keywords when BOOST or SUPPRESS modes are active. This variance allows for slight adjustments around the maximum or minimum weight, adding a layer of randomness and preventing uniformity in keyword weighting. This parameter is important for creating a more dynamic and varied weighting effect.

Weight Lifter 🏋️‍♀️ Output Parameters:

weighted_tags

The weighted_tags output parameter provides the final list of tags with their applied weights. Each tag is presented with its corresponding weight, formatted with exact whitespace and newlines preserved. This output is crucial for understanding the impact of the weight application process and for further use in creative projects.

Weight Lifter 🏋️‍♀️ Usage Tips:

• Experiment with different mode settings to achieve the desired weight distribution effect, such as using GRADIENT for smooth transitions or BURST for focused clusters.
• Utilize keyword_mode to emphasize or de-emphasize specific keywords, enhancing their impact or reducing their influence as needed.
• Adjust min_weight and max_weight to control the range of weights applied, ensuring that no tag is overly diminished or excessively emphasized.

Weight Lifter 🏋️‍♀️ Common Errors and Solutions:

"Invalid mode selected"

• Explanation: This error occurs when an unsupported mode is chosen for the mode parameter.
• Solution: Ensure that the mode parameter is set to one of the supported options: RANDOM, GRADIENT, BURST, or NOISE.

"Keyword mode not recognized"

• Explanation: This error indicates that an unrecognized value was provided for the keyword_mode parameter.
• Solution: Verify that the keyword_mode is set to a valid option: NONE, ONLY, IGNORE, BOOST, or SUPPRESS.

"Weight out of bounds"

• Explanation: This error happens when a calculated weight exceeds the specified min_weight or max_weight limits.
• Solution: Check the min_weight and max_weight settings to ensure they are correctly defined and adjust them if necessary.