[2884]: Fix cross_attention_kwargs in StableDiffusionImg2ImgPipeline

[nipunjindal]

Issue: #2884
Fixing cross_attention_kwargs in img2img pipeline.
Please use below code to reproduce the bug:

import torch
import requests
from PIL import Image
from io import BytesIO

from diffusers import StableDiffusionImg2ImgPipeline

# load the pipeline
device = "mps"
model_path = "sayakpaul/sd-model-finetuned-lora-t4"
pipe = StableDiffusionImg2ImgPipeline.from_pretrained("CompVis/stable-diffusion-v1-4", torch_dtype=torch.float16).to(
    device
)
pipe.unet.load_attn_procs(model_path)

# let's download an initial image
url = "https://raw.githubusercontent.com/CompVis/stable-diffusion/main/assets/stable-samples/img2img/sketch-mountains-input.jpg"

response = requests.get(url)
init_image = Image.open(BytesIO(response.content)).convert("RGB")
init_image.thumbnail((768, 768))

prompt = "A fantasy landscape, trending on artstation"

images = pipe(prompt=prompt, image=init_image, strength=0.75, guidance_scale=7.5, cross_attention_kwargs={"scale": 0.5}).images

display(images[0])

[nipunjindal]

Hello @sayakpaul, @patrickvonplaten,

Please review the PR. Thank you!

[sayakpaul]

Well done. Thanks a lot for adding this.

[HuggingFaceDocBuilderDev]

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

[patrickvonplaten]

Thanks!