Waterfront XL 0.8 Update Release Notes
We are thrilled to announce the release of Waterfront XL version 0.8! This update brings significant enhancements and features to elevate your creative experience.
Stronger Anatomy: The new version offers improved anatomical accuracy, resulting in more lifelike and detailed images.
Better Character Recognition: More characters can now be recognized by their names without requiring additional LoRA models.
More Style Variations: Enhanced style control with new tags. Use year tags (2020s, 2015s, 2010s, 2000s), style tags (chibi, pixel art, anime screencap, etc.), or artist tags (some_artist) to achieve the desired style.
We recommend using danbooru-like tags for this model and ordering the tags in the format of <artists> <characters> <everything else> <rating/year>.
The model is sensitive to higher prompt weights. Using weights like ("a tag":1.5) or (((((amazing details))))) could potentially distort the image.
- Spacing: Use spaces for most inputs (looking at viewer), and use underscores if found in certain artist names (some_artist).
- Brackets: Use \( and \) for literal brackets to avoid unintended weight adjustments. For example: Janne D'arc\\(FGO\\)
Some style tags may require adjusted weights to achieve the desired effect:
- Higher Weight: Use double parentheses (( )) around prompts that need more emphasis (e.g., ((some_style))).
- Lower Weight: Use double square brackets [[ ]] for less emphasis.
- Quality Tags: best quality, good quality, average quality, bad quality, worst quality.
- Year Tags: 2020s, 2015s, 2010s, 2000s.
- Resolution Tags: absurdres, highres, lowres.
- Rating Tags: general, questionable, sensitive, explicit.
For this model, we add best quality, 2020s, highres to the end of your prompt by default for ease of use. Additional quality tags are usually not needed as they're added by default, and adding too many could potentially overwhelm the model. Also we do not recommand adding LoRA when using this model.
We encourage you to explore these new features and see the improvements in Waterfront XL 0.8. Experiment with different styles and prompts to discover the full potential of the model. Your feedback is crucial in helping us refine and perfect the model. Enjoy creating with the latest version!
Key Features
Stronger Anatomy: The new version offers improved anatomical accuracy, resulting in more lifelike and detailed images.
Better Character Recognition: More characters can now be recognized by their names without requiring additional LoRA models.
More Style Variations: Enhanced style control with new tags. Use year tags (2020s, 2015s, 2010s, 2000s), style tags (chibi, pixel art, anime screencap, etc.), or artist tags (some_artist) to achieve the desired style.
Model Details
We recommend using danbooru-like tags for this model and ordering the tags in the format of <artists> <characters> <everything else> <rating/year>.
The model is sensitive to higher prompt weights. Using weights like ("a tag":1.5) or (((((amazing details))))) could potentially distort the image.
Input Format
- Spacing: Use spaces for most inputs (looking at viewer), and use underscores if found in certain artist names (some_artist).
- Brackets: Use \( and \) for literal brackets to avoid unintended weight adjustments. For example: Janne D'arc\\(FGO\\)
Prompt Weighting
Some style tags may require adjusted weights to achieve the desired effect:
- Higher Weight: Use double parentheses (( )) around prompts that need more emphasis (e.g., ((some_style))).
- Lower Weight: Use double square brackets [[ ]] for less emphasis.
Available Special Tags
- Quality Tags: best quality, good quality, average quality, bad quality, worst quality.
- Year Tags: 2020s, 2015s, 2010s, 2000s.
- Resolution Tags: absurdres, highres, lowres.
- Rating Tags: general, questionable, sensitive, explicit.
For this model, we add best quality, 2020s, highres to the end of your prompt by default for ease of use. Additional quality tags are usually not needed as they're added by default, and adding too many could potentially overwhelm the model. Also we do not recommand adding LoRA when using this model.
We encourage you to explore these new features and see the improvements in Waterfront XL 0.8. Experiment with different styles and prompts to discover the full potential of the model. Your feedback is crucial in helping us refine and perfect the model. Enjoy creating with the latest version!
Updated on: 26/03/2025
Thank you!