[Waterfront XL Demo - A New PixAI Exclusive Model]
Introducing a new PixAI exclusive model in the works: Waterfront XL
This model is a temporary demo of a PixAI-trained model, currently about 60% complete. There are some kown issues that will be improved in the next version, Waterfront XL posesses some great potential that we are very excited about. We encourage everyone to use it and provide feedback to help us train an even better model in the future.
You can get a taste of Waterfront XL on our site today. This model excels at stylization for all anime-styled art.
Please note that the prompt and settings for this model are different from what you're used to. Follow these instructions for the best results:
- Prompts: This model will automatically add the quality tags for you(best quality, absurdres, masterpiece, etc.) YOU DON'T HAVE TO ADD ANY OF THESE! Please just describe the basic description of the image in your prompts
- Character Tags: This model only supports character tags that pre-exist on Danbooru. It does not recognize other character names or names written in other formats.
- ControlNet: We do not recommend using ControlNet with this model.
- Sampling Methods: Supports all sampling methods. Euler with 28 steps is recommended.
- Hi-Res or Upscale: Recommended at 1.5x with 0.6 denoise.
This model is a temporary demo of a PixAI-trained model, currently about 60% complete. There are some kown issues that will be improved in the next version, Waterfront XL posesses some great potential that we are very excited about. We encourage everyone to use it and provide feedback to help us train an even better model in the future.
You can get a taste of Waterfront XL on our site today. This model excels at stylization for all anime-styled art.
Please note that the prompt and settings for this model are different from what you're used to. Follow these instructions for the best results:
- Prompts: This model will automatically add the quality tags for you(best quality, absurdres, masterpiece, etc.) YOU DON'T HAVE TO ADD ANY OF THESE! Please just describe the basic description of the image in your prompts
- Character Tags: This model only supports character tags that pre-exist on Danbooru. It does not recognize other character names or names written in other formats.
- ControlNet: We do not recommend using ControlNet with this model.
- Sampling Methods: Supports all sampling methods. Euler with 28 steps is recommended.
- Hi-Res or Upscale: Recommended at 1.5x with 0.6 denoise.
Updated on: 26/03/2025
Thank you!