How to use
This is a cleaned up version of the Realistic Vision 5.1 model. Generated fp16
versions and extracted the inpainting model. This can be found under the branch realistic-vision-5_1-inpainting
.
from diffusers import StableDiffusionPipeline, StableDiffusionInpaintPipeline
pipeline = StableDiffusionPipeline.from_pretrained(
"holwech/realistic-vision-optimized", variant="fp16", torch_dtype=torch.float16
)
# OR
pipeline = StableDiffusionInpaintPipeline.from_pretrained(
"holwech/realistic-vision-optimized", revision="realistic-vision-5_1-inpainting", variant="fp16", torch_dtype=torch.float16
)