Type | Upscaler |
Stats | 151,838 0 |
Reviews | (3,480) |
Published | Jul 25, 2023 |
Base Model | |
Training | Steps: 1 Epochs: 1 |
Usage Tips | Clip Skip: 1 |
Hash | AutoV2 A5812231FC |




Check my EXCLUSIVE models on Mage.Space: AniMage PXL • AniReal PXL • Lucid Dream • AniMage SD1.5 • Realistic Portrait
SDXL - Pony: AniVerse PXL • AniMerge PXL • AniToon PXL • AniMics PXL • AniVerse XL
SD1.5: AniVerse • AniThing • AniMerge • AniMesh • AniToon • AniMics
Also in Collaboration with Shakker.ai
This upscaler is not mine, all the credit go to: Kim2091
Official WiKi Upscaler page: Here
License of use it: Here
HOW TO INSTALL:
Rename the file from:
4x-UltraSharp.pt to: 4x-UltraSharp.pth
Copy the file 4x-UltraSharp.pth inside the folder: "\YOUR~STABLE~DIFFUSION~FOLDER\models\ESRGAN\")
Restart you Stable Diffusion
Now in "Extra Tab" you got the new upscaler
Discussion
Your file's name is 4xUltrasharp_4xUltrasharpV10.pt
I can't use it at hires.fix , only extra's upscale works.
I downloaded again from Kim's website. It is 4x-UltraSharp.pth Different extension. It's PTH, not PT
PTH works fine, upscale and hires.fix works fine.
Btw, I checked sha256s too. Kim's and yours are both A5812231FC936B42AF08A5EDBA784195495D303D5B3248C24489EF0C4021FE01
I ran into a problem with trying to use this upscaler:
```
Error occurred when executing UpscaleModelLoader: Weights only load failed. Re-running torch.load with weights_only set to False will likely succeed, but it can result in arbitrary code execution.Do it only if you get the file from a trusted source. WeightsUnpickler error: Unsupported operand 80
```
So this means the upscaler is demanding to be run in such a state that it can run anything on your computer. Kim2091 is the original author of this model and you can find this model here: https://huggingface.co/Kim2091/UltraSharp/tree/main
I have taken SHA256s of the two files:
Samael1976:
sha256-pYEiMfyTa0KvCKXtunhBlUldMD1bMkjCRInvDEAh/gE=
Kim2091:
sha256-weKbF1v8Yi1+8r/eagg6FWJ5QFdFmmTLXZlJFQdFiIU=
They are different, which means this model is not the exact same copy of Kim2091's. This could just be an older version of Kim's, or it could be something fishy is going on. I try to keep an open mind and so I went to verify.
Samael's copy is about 10 months old at time of writing (2024-04-05) but Kim's most recent version is 3 months old. Perhaps it just needs an update. Unfortunately I cannot see history in the git repository that goes back that far to verify. If updated, I do expect those SHAs should match.
Using the latest version did not work around the issue I had above, so maybe it's okay but it's still hard to do full audits on these things. I have not done a full audit here.
What's the difference between the original UltraSharp at https://openmodeldb.info/models/4x-UltraSharp and your UltraSharp.
how to use this upscaler on Foocus Stable diffusion? does it work on focus? please tell how to make it work