forked from Coyote-A/ultimate-upscale-for-automatic1111
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
2 changed files
with
26 additions
and
73 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,87 +1,40 @@ | ||
| # Ultimate SD upscaler for automatic1111 web-ui | ||
| # Ultimate SD Upscale extension for [AUTOMATIC1111 Stable Diffusion web UI](https://github.com/AUTOMATIC1111/stable-diffusion-webui) | ||
| Now you have the opportunity to use a large denoise (0.3-0.5) and not spawn many artifacts. Works on any video card, since you can use a 512x512 tile size and the image will converge. | ||
|
|
||
| # How to install | ||
| 1. Go to "Extensions" tab | ||
| 2. Select "Install from URL" | ||
| 3. Put https://github.com/Coyote-A/ultimate-upscale-for-automatic1111 into "URL for extension's git repository" field | ||
| 4. On "Installed" tab click "Apply and restart UI" | ||
| 5. ??? | ||
| 6. Profit | ||
|
|
||
| # How to use | ||
| Send image to img2img, in "Script" field select "Ultimate SD upscale". The width and height of the future image is selected through the default slider. | ||
|
|
||
| Set **Denoising strength** in default settings for redraw pass. | ||
|
|
||
|
|
||
|  | ||
|
|
||
| Enable redraw if it disabled. Select tile size, change padding and mask blur. | ||
|
|
||
| Select upscaler. If you don't want to use the upscaler it's ok, but the image will be blurry. | ||
|
|
||
| If a tile grid is visible on your image, then you can run the image through "Sems fix". Just send it to img2img, enable them and disable redraw. Set upscaler to None. | ||
|
|
||
| # Some recommendations | ||
|
|
||
| For 4k image use: | ||
| * **Tile size**: 768 | ||
| * **Mask blur**: 20 | ||
| * **Padding**: 55 | ||
| * **Denoising strength**: 0.35-0.45 | ||
|
|
||
| For 2k image you can use (you can try it on 4k, but result have many mutations): | ||
| * **Tile size**: 512 | ||
| * **Mask blur**: 16 | ||
| * **Padding**: 32 | ||
| * **Denoising strength**: 0.35-0.45 | ||
|
|
||
| Or the same settings as for 4k | ||
|
|
||
| Recommended upscaler: R-ESRGAN 4x+ | ||
|
|
||
| There are 2 types of seam fix: **half tile offset pass** and **bands pass** | ||
|
|
||
| Do not use seam fixes if image haven't visible grid, it's just another redraw passes. | ||
|
|
||
| ### Half tile offset pass | ||
| It adds 2 passes like redraw pass, but with a half-tile offset. One pass for rows with vertical gradient mask and one pass for columns with horizontal gradient mask. | ||
|
|
||
| This pass covers bigger area than bands and mostly produces better result, but require more time. | ||
|
|
||
| ### Bands pass | ||
| It adds passes on just seams (rows and columns) and covers small area around them (*width* in ui). It requires less time than offset pass. | ||
|
|
||
| ### Half tile offset + intersections pass | ||
| It runs Half tile offset pass and then run extra pass on intersections with radial gradient mask | ||
|
|
||
| # How to enable 4096 img width and height | ||
| 1. Select **Custom size** in **Target size type** | ||
| 2. Set **custom width** and **custom height** to 4096 | ||
| # Instructions | ||
| All instructions can be found on the project's [wiki](https://github.com/Coyote-A/ultimate-upscale-for-automatic1111/wiki). | ||
|
|
||
| # Examples | ||
|
|
||
| Original image | ||
| <details> | ||
| <summary>E1</summary> | ||
| Original image | ||
|
|
||
|  | ||
|  | ||
|
|
||
| 2k upscaled. **Tile size**: 512, **Padding**: 32, **Mask blur**: 16, **Denoise**: 0.4 | ||
|  | ||
| 2k upscaled. **Tile size**: 512, **Padding**: 32, **Mask blur**: 16, **Denoise**: 0.4 | ||
|  | ||
| </details> | ||
|
|
||
| Original image | ||
| <details> | ||
| <summary>E2</summary> | ||
| Original image | ||
|
|
||
|  | ||
|  | ||
|
|
||
| 2k upscaled. **Tile size**: 768, **Padding**: 55, **Mask blur**: 20, **Denoise**: 0.35 | ||
|  | ||
| 2k upscaled. **Tile size**: 768, **Padding**: 55, **Mask blur**: 20, **Denoise**: 0.35 | ||
|  | ||
|
|
||
| 4k upscaled. **Tile size**: 768, **Padding**: 55, **Mask blur**: 20, **Denoise**: 0.35 | ||
|  | ||
| 4k upscaled. **Tile size**: 768, **Padding**: 55, **Mask blur**: 20, **Denoise**: 0.35 | ||
|  | ||
| </details> | ||
|
|
||
| Original image | ||
| <details> | ||
| <summary>E3</summary> | ||
| Original image | ||
|
|
||
|  | ||
|  | ||
|
|
||
| 4k upscaled. **Tile size**: 768, **Padding**: 55, **Mask blur**: 20, **Denoise**: 0.4 | ||
|  | ||
| 4k upscaled. **Tile size**: 768, **Padding**: 55, **Mask blur**: 20, **Denoise**: 0.4 | ||
|  | ||
| </details> |