when will Image Upscaling with Stable Diffusion function be available? about stablediffusion

[ThibaultLSDC]

Hi, sorry I deleted my answer 'cause the link didn't work in github for some reason. (?) The model checkpoints are download-able in the Hugginface model card.
Anyways your error is because you're calling both scripts/gradio/inpainting.py and scripts/gradio/superresolution.py. You should run:
!python scripts/gradio/superresolution.py configs/stable-diffusion/x4-upscaling.yaml x4-upscaler-ema.ckpt
for upscaling.
I ran into the same issue and I guess there might have been a mistake in the README yesterday ?
Hope this helps

[mule-engineer13]

Thank you for answering！
I'll try！

[lucellent]

Hi, sorry I deleted my answer 'cause the link didn't work in github for some reason. (?) The model checkpoints are download-able in the Hugginface model card. Anyways your error is because you're calling both scripts/gradio/inpainting.py and scripts/gradio/superresolution.py. You should run: !python scripts/gradio/superresolution.py configs/stable-diffusion/x4-upscaling.yaml x4-upscaler-ema.ckpt for upscaling. I ran into the same issue and I guess there might have been a mistake in the README yesterday ? Hope this helps

Is it possible to use the upscaler on our own images, and not have to first generate SD images?

[mule-engineer13]

Thank you for answering.
I want to upscale a local image. Is upscaling possible even if the image is not generated by SD?

[mule-engineer13]

Using the code you gave me below, I was able to start inference.
!python scripts/gradio/superresolution.py configs/stable-diffusion/x4-upscaling.yaml x4-upscaler-ema.ckpt

However, even after 30 minutes or more, it does not proceed from the following display.

●log

No module 'xformers'. Proceeding without it.
LatentUpscaleDiffusion: Running in v-prediction mode
DiffusionWrapper has 473.40 M params.
making attention of type 'vanilla' with 512 in_channels
Working with z of shape (1, 4, 64, 64) = 16384 dimensions.
making attention of type 'vanilla' with 512 in_channels
Downloading: 100% 3.94G/3.94G [01:11<00:00, 55.5MB/s]
/usr/local/lib/python3.7/dist-packages/gradio/deprecation.py:44: UserWarning: You have unused kwarg parameters in Number, please remove them: {'min': 0.0, 'max': 1.0}
  f"You have unused kwarg parameters in {cls}, please remove them: {kwargs}"
/usr/local/lib/python3.7/dist-packages/gradio/deprecation.py:44: UserWarning: You have unused kwarg parameters in Number, please remove them: {'min': 0, 'max': 350, 'step': 1}
  f"You have unused kwarg parameters in {cls}, please remove them: {kwargs}"
Running on local URL:  http://127.0.0.1:7860/

To create a public link, set `share=True` in `launch()`.

●gpu version(googlecolab free version gpu)

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Sun_Feb_14_21:12:58_PST_2021
Cuda compilation tools, release 11.2, V11.2.152
Build cuda_11.2.r11.2/compiler.29618528_0

●pytorch version

open-clip-torch               2.7.0
pytorch-lightning             1.4.2
torch                         1.10.0
torchaudio                    0.10.0
torchmetrics                  0.6.0
torchsummary                  1.5.1
torchtext                     0.11.0
torchvision                   0.10.1

Does it take this long?
If you know any improvement measures, could you tell me?

[kongfanyu06]

Using the code you gave me below, I was able to start inference. !python scripts/gradio/superresolution.py configs/stable-diffusion/x4-upscaling.yaml x4-upscaler-ema.ckpt

However, even after 30 minutes or more, it does not proceed from the following display.

●log

No module 'xformers'. Proceeding without it.
LatentUpscaleDiffusion: Running in v-prediction mode
DiffusionWrapper has 473.40 M params.
making attention of type 'vanilla' with 512 in_channels
Working with z of shape (1, 4, 64, 64) = 16384 dimensions.
making attention of type 'vanilla' with 512 in_channels
Downloading: 100% 3.94G/3.94G [01:11<00:00, 55.5MB/s]
/usr/local/lib/python3.7/dist-packages/gradio/deprecation.py:44: UserWarning: You have unused kwarg parameters in Number, please remove them: {'min': 0.0, 'max': 1.0}
  f"You have unused kwarg parameters in {cls}, please remove them: {kwargs}"
/usr/local/lib/python3.7/dist-packages/gradio/deprecation.py:44: UserWarning: You have unused kwarg parameters in Number, please remove them: {'min': 0, 'max': 350, 'step': 1}
  f"You have unused kwarg parameters in {cls}, please remove them: {kwargs}"
Running on local URL:  http://127.0.0.1:7860/

To create a public link, set `share=True` in `launch()`.

●gpu version(googlecolab free version gpu)

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Sun_Feb_14_21:12:58_PST_2021
Cuda compilation tools, release 11.2, V11.2.152
Build cuda_11.2.r11.2/compiler.29618528_0

●pytorch version

open-clip-torch               2.7.0
pytorch-lightning             1.4.2
torch                         1.10.0
torchaudio                    0.10.0
torchmetrics                  0.6.0
torchsummary                  1.5.1
torchtext                     0.11.0
torchvision                   0.10.1

Does it take this long? If you know any improvement measures, could you tell me?
Open this url: http://127.0.0.1:7860/ on your browser, then choose a image to run.

[mule-engineer13]

Access is denied when opening http://127.0.0.1:7860/.
A screen with ERR_CONNECTION_REFUSED is displayed.

[mule-engineer13]

I have an additional question.
I tried the demo below. Is this stable-diffusion 2.0?
https://huggingface.co/spaces/stabilityai/stable-diffusion-upscaler

[mule-engineer13]

Access is denied when opening http://127.0.0.1:7860/.
A screen with ERR_CONNECTION_REFUSED is displayed.

I was able to get this to work using a public URL.
Thank you.

[Qingcsai]

@mule-engineer13 Hi, how did you get the http://127.0.0.1:7860/ work using a public URL, I also show this but my access is denied when opending it.