Rainfield - Horror Scene

  This lora is based on training a realistic large model as the base model, using more than 50 materials for training, you can try any words you can think of, and enjoy the joy of drawing cards to your heart's content.

Recommended raw image size: 512*1024

Recommended weight: 1.0.

Recommended steps: 30.

VAE: Not required

Highly recommend opening the high-score repair Hires.fix

Enlargement algorithm: 4X-UltraSharp, I only use this one, whether it's real people or two-dimensional, everyone can choose their own.

For character generation, it is recommended to use the ADetailer facial optimization plugin

Feel free to try other options,

Requesting all the great masters to give me feedback! Thank you!