sukkritsharma's picture
Update README.md
a624610 verified
|
raw
history blame
970 Bytes
metadata
license: creativeml-openrail-m

This is an inpainting model, which has been converted from the realisticVisionV51_v51VAE-inpainting.

How to load the model :

model_id = "sukkritsharma/realisticVisionV51_v51VAE-inpainting-onnx-fp16"
pipe = OnnxStableDiffusionInpaintPipeline.from_pretrained(model_id, 
                                                              revision='fp16',
                                                              provider="CUDAExecutionProvider",
                                                             )

How to run inference :

img = pipe(prompt=prompt,
           image=init_image,
           mask_image=mask_image,
           num_inference_steps=30).images[0]