update the readme to fix the image-to-image example

- add `import torch`
- include `pipe.to("cuda")`

README.md

@@ -95,8 +95,10 @@ below.
 ```py
 from diffusers import AutoPipelineForImage2Image
 from diffusers.utils import load_image
+import torch
 
 pipe = AutoPipelineForImage2Image.from_pretrained("stabilityai/sdxl-turbo", torch_dtype=torch.float16, variant="fp16")
+pipe.to("cuda")
 
 init_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png").resize((512, 512))