Using an image as visual prompt for generation

[TheDenk]

Hey :)

I've made some changes for using visual prompt image. It works nice with text prompt combinations and without them.
I tried to keep the style of writing code like yours.
Changes:

1. Little changes in main pipelines.
2. Code examples and image example were added in readme.
3. Gradio demo was changed for load additional image.
4. Additional file was added for infer with image prompt.

Examples:

[haofanwang]

Interesting!

[haofanwang]

I tested it locally and found that it might be also interesting to set the lighting direction as well.

But I think I would be better to add it as an individual pipeline_stable_diffusion_xl_instantid_img2img.py. We want to make the original pipeline as simple as possible. Thanks.

[TheDenk]

I separated pipelines and gradio apps to original and img2img.
All changes have been tested.

@haofanwang

[ciyuanjing]

我在本地对其进行了测试，发现设置照明方向也可能很有趣。

但我认为我最好将其作为个人添加。我们希望使原始管道尽可能简单。谢谢。`pipeline_stable_diffusion_xl_instantid_img2img.py`

perfect