refactor kandinsky img2img for ddpm

[yiyixuxu]

this PR refactor the img2img so that it's compatible with ddpm scheduler

fromdiffusersimportKandinskyImg2ImgPipeline, KandinskyPriorPipeline, DDPMSchedulerfromPILimportImageimporttorchREPO_PRIOR="kandinsky-community/kandinsky-2-1-prior"REPO="kandinsky-community/kandinsky-2-1"model_dtype=torch.float16prompt="A red cartoon frog, 4k"# create prior pipe_prior=KandinskyPriorPipeline.from_pretrained(REPO_PRIOR, torch_dtype=model_dtype)
pipe_prior.to("cuda")
# create img2img pipelinescheduler=DDPMScheduler.from_pretrained(REPO, subfolder="ddpm_scheduler")
pipe=KandinskyImg2ImgPipeline.from_pretrained(REPO, scheduler=scheduler, torch_dtype=model_dtype)
pipe.to("cuda")
# use prior to generate image_emb based on our promptgenerator=torch.Generator(device='cuda').manual_seed(0)
image_emb, zero_image_emb=pipe_prior(prompt, num_inference_steps=25, generator=generator,).to_tuple()
init_image=Image.open("frog.png")
generator=torch.Generator(device='cuda').manual_seed(0)
out=pipe(
prompt=prompt, image=init_image, height=768, width=768, num_inference_steps=100, generator=generator, image_embeds=image_emb, negative_image_embeds=zero_image_emb, strength=0.2,) out.images[0].save("frog_out.png")

[HuggingFaceDocBuilderDev]

The documentation is not available anymore as the PR was closed or merged.

[patrickvonplaten]

Let's maybe add a test here as well no?

[patrickvonplaten]

@yiyixuxu let's try to get this in :-)

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[patrickvonplaten]

@yiyixuxu is this still open?