Iopaint

- Redo: Now you can `Redo` the inpainting result
<img width="297" alt="image" src="https://user-images.githubusercontent.com/3998421/176456044-199349c4-eb1f-477c-a37a-cb0fb6bf5b0b.png">


- Gradually Inpainting(enabled by default): Perform inpainting on the last inpainting result, this makes it possible to further improve the inpainting result. It also helps to keep unmasked previous inpainting regions unchanged

<img width="708" alt="image" src="https://user-images.githubusercontent.com/3998421/176455800-777d3933-2e64-453f-869e-54f6c9c6fc8b.png">

- `Start New` change to `Upload New`: upload new file at editor page
<img width="188" alt="image" src="https://user-images.githubusercontent.com/3998421/176459988-dc664dcc-7fbc-4bc1-91b9-3c87ffd4d746.png">

latent-diffusion model add [plms](https://arxiv.org/abs/2202.09778) sampler. In general plms can get better results with fewer steps.

Original Image

ddim sampler steps=5

plms sampler steps=5

- bug fix: fix sizeLimit reset after inpainting.
- ldm model fp16 support, use less GPU memory. Benchmark on 3090

|Image size| GPU memory|
|-----------|-------------|
| 720x519 | 6147mb|
|1080x778|17499mb|

- Manual inpainting mode: It is convenient to use when there are many target objects
- In manual mode you can undo strokes

<img width="480" alt="image" src="https://user-images.githubusercontent.com/3998421/164982076-aebb962e-b476-4902-9dbf-3ccf6762e423.png">

https://user-images.githubusercontent.com/3998421/164982322-4f0baa60-3b38-4ec7-9ea3-36bb0d152117.mp4

- zoom bug fix
- editor tool panel CSS style fix

- Install by pip: pip install lama-cleaner
- Settings: now you can change model without restart the server
<img width="786" alt="image" src="https://user-images.githubusercontent.com/3998421/163836640-f16cfe9c-759f-4962-9454-b2779cbdcf49.png">