Eve Valentine

1: completely redraw face or hands / 0: no effect on output images

The number of last layers of CLIP network to skip

Enhance/diminish detail while keeping the overall style/character

Disable automatically adding suggested prompt modification. Built-in LoRAs and trigger words will remain.

Embedding file (textural inversion). Apply by writing the following in prompt or negative prompt: (FILE_NAME_WITHOUT_EXTENSION:MAGNITUDE)

Base image that the output should be generated from. This is useful when you want to add some detail to input_image. For example, if prompt is "sunglasses" and input_image has a man, there is the man wearing sunglasses in the output.

How differ the output is from input_image. Used only when input_image is given.

LoRA file. Apply by writing the following in prompt: <lora:FILE_NAME_WITHOUT_EXTENSION:MAGNITUDE>

Strength of applying reference_depth_image. Used only when reference_depth_image is given.

Image with which the output should share identity (e.g. face of a person or type of a dog)

Strength of applying reference_image. Used only when reference_image is given.

Strength of applying reference_pose_image. Used only when reference_pose_image is given.

Same seed with the same prompt generates the same image. Set as -1 to randomize output.

real-dream-sdxl