Lama-cleaner

1. Remove the landing page, now you can direct drag the image to the editor

https://user-images.githubusercontent.com/3998421/194713446-c22e8320-1f12-4910-814b-ae2b0f1a3ade.mov


https://user-images.githubusercontent.com/3998421/194713757-21b637a1-fc5d-408a-987b-337627a3f33b.mov




2. Use `ctrl+v` to paste an image to the editor

https://user-images.githubusercontent.com/3998421/194713285-0dd5b826-46f7-44d8-b228-8d395d68fdc6.mov

Add three args for `sd1.4`

| Name | Description | Default |
|-------------------|-------------------------------------------------------------------------------------------------------------------------------| -------- |
| --sd-run-local | Once the model are downloaded, you can pass this arg and remove `--hf_access_token` | |
| --sd-disable-nsfw | Disable stable-diffusion NSFW checker. | |
| --sd-cpu-textencoder | Always run stable-diffusion TextEncoder model on CPU. | |

fix sd inpainting bug, currently diffusers only support mod 64 forward: https://github.com/huggingface/diffusers/pull/505

Add `ctrl/cmd + c` shortcut to copy inpainting result

- Add OpenCV2 inpainting method: Suitable for simple backgrounds
- Allow drag image on the full page to upload a new image

To use sd1.4 you need to get an access token from https://huggingface.co/docs/hub/security-tokens


https://user-images.githubusercontent.com/3998421/192548839-9d0cbfc5-6bf7-4a5a-b556-446878696bdf.mp4



You need the right prompt and a lot of luck to get good results. Here are some of my test results