he means run the notebook

jupiter notebook

is the command
you need to install it pip/conda

pip/conda

Please tell me the exact commands i need to use:

Also, i have Anaconda/Conda installed already:

Thx in advance!

he means run the notebook

jupiter notebook

is the command you need to install it pip/conda

well, it doesn't run...

Just type those commands one by one

conda create -n sdinf python=3.10 -y
conda activate sdinf
pip install notebook
jupyter notebook

Just type those commands one by one

conda create -n sdinf python=3.10 -y
conda activate sdinf
pip install notebook
jupyter notebook

Big thx, i'm in.

Now, how i get to the Canvas to start Outpainting?

Errors in every cell i try to RUN...

For Windows, you may need to install anaconda or miniconda at first and then:

conda create -n sd-inf python=3.10
conda activate sd-inf
conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch
conda install scipy
conda install -c conda-forge jupyterlab 
conda install -c conda-forge ipycanvas
conda install -c conda-forge diffusers transformers ftfy
pip install opencv-python
git clone https://github.com/lkwq007/stablediffusion-infinity
cd stablediffusion-infinity
huggingface-cli login
jupyter lab
# and open stablediffusion_infinity_windows.ipynb

Note that the notebook has not been tested on Windows, so I am not sure whether it would work well.

@lkwq007 Hi, i'm not the OP of this thread, but I'm attempting to utilize this on Windows. All of this seems to have worked, but I needed to use conda install opencv instead of pip install opencv-python.

Sadly i'm running into this issue in the upload image cell in the stablediffusion_infinity_windows.ipynb:

---

AttributeError Traceback (most recent call last)
Cell In [7], line 14, in start_func(btn)
12 uploader_output.clear_output()
13 if len(uploader.value)>0:
---> 14 keys=list(uploader.value.keys())
15 val=uploader.value[keys[-1]]["content"]
16 pil=Image.open(io.BytesIO(val))

AttributeError: 'tuple' object has no attribute 'keys'

Kim - running into same "tuple" error AT TIMES
the uploader is confusing, you can select the image but when confirming says no image uploaded
have tried to crop the images to ingest to 512/512 in case resizer was a problem

if useful

• instruction don't work on M1 macs (for obvious ARM reasons). Windows works
• make sure you are using a token from an account on hugging face where you approved the Terms and conditions

Help on how to upload an image which is placed in the working area would be appreciated

Thank you for the reply. I am on Windows, and I've now verified that I'm logged in. Sadly it's still yielding the same error. Inpainting itself actually does function, it's just uploading an image that is broken as you specified.

@lkwq007 Hi, i'm not the OP of this thread, but I'm attempting to utilize this on Windows. All of this seems to have worked, but I needed to use conda install opencv instead of pip install opencv-python.

Sadly i'm running into this issue in the upload image cell in the stablediffusion_infinity_windows.ipynb:

AttributeError Traceback (most recent call last)
Cell In [7], line 14, in start_func(btn)
12 uploader_output.clear_output()
13 if len(uploader.value)>0:
---> 14 keys=list(uploader.value.keys())
15 val=uploader.value[keys[-1]]["content"]
16 pil=Image.open(io.BytesIO(val))
AttributeError: 'tuple' object has no attribute 'keys'

Seems to be ipywidgets issue, how about conda install -c conda-forge ipywidgets=7.7.1 nodejs?

conda install -c conda-forge ipywidgets=7.7.1 nodejs

Thank you, that fixed the issue with uploading images. Unfortunately a new error has occurred in the following cell:

---

NameError Traceback (most recent call last)
Cell In [6], line 4
2 upload_button.disabled=True
3 display_lst=base.display()
----> 4 display_lst.insert(-1,widgets.HBox([resize_check,step_input,guidance_input,strength_slider]))
5 for item in display_lst:
6 display(item)

NameError: name 'resize_check' is not defined

EDIT: Never mind, my apologies. A cell failed to execute and I overlooked it. Thank you for the help!

but how to use local pretrained models? i have no account to hugginface, is possiblo to run offline with downloaded pretrained models? (like automatic111 if you know)

you can open a free account on HF. you'll need an access token

conda install -c conda-forge ipywidgets=7.7.1 nodejs

Thanks for the command!
after I run that, start again and get this error in the 'upload an image' and 'have fun here' cells:

what can I do? thanks for your time!!

after I run that, start again and get this error in the 'upload an image' and 'have fun here' cells:

@spinferno Try installing jupyterlab_widgets 1.1.1, that seems to work for me

after I run that, start again and get this error in the 'upload an image' and 'have fun here' cells:

@spinferno Try installing jupyterlab_widgets 1.1.1, that seems to work for me

thanks so much for your help!
as per the latest commit:
7a2a1d9

I ran these:

conda install -c conda-forge ipywidgets=7.7.1 nodejs
conda install -c conda-forge jupyterlab_widgets=1.1.1 nodejs
conda install -c conda-forge widgetsnbextension=3.6.1 nodejs

however same exception is thrown, sadly.

🙂 You may want to make a fresh clone of the repo and try a new environment from scratch. I ran into similar issues, doing this seems to fix it. Maybe having the older dependencies and running it at all somehow introduces permanent issues to that copy?

I hope you get this figured out. I made an issue #12 that documented many of the issues I ran into earlier today. You may want to check that out for any solutions.

i just wiped my env and stood up everything again. somehow the exception remains and now all the UI buttons are still broken, include the generate button, how else can i fix things?
this is what I did:

remove installation
conda env remove -n sd-inf

rebuild
conda create -n sd-inf python=3.10 -y
conda activate sd-inf
conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch -y
conda install scipy -y
conda install -c conda-forge jupyterlab -y
conda install -c conda-forge ipywidgets=7.7.1 -y
conda install -c conda-forge ipycanvas -y
conda install -c conda-forge diffusers transformers ftfy -y
pip install opencv-python
conda activate sd-inf
jupyter lab
tested, shut down

fix dependencies
conda install -c conda-forge ipywidgets=7.7.1 nodejs
conda install -c conda-forge jupyterlab_widgets=1.1.1 nodejs
conda install -c conda-forge widgetsnbextension=3.6.1 nodejs
jupyter lab
tested, still getting runtime errors

runtime errors

for the file upload widget:
[Open Browser Console for more detailed log - Double click to close this message]
Failed to load model class 'LabelModel' from module '@jupyter-widgets/controls'
Error: Module @jupyter-widgets/controls, version ^1.5.0 is not registered, however, 2.0.0 is
at f.loadClass (http://localhost:8888/lab/extensions/@jupyter-widgets/jupyterlab-manager/static/134.bcbea9feb6e7c4da7530.js?v=bcbea9feb6e7c4da7530:1:74977)
at f.loadModelClass (http://localhost:8888/lab/extensions/@jupyter-widgets/jupyterlab-manager/static/150.3e1e5adfd821b9b96340.js?v=3e1e5adfd821b9b96340:1:10729)

for the uplaod confirm button:
[Open Browser Console for more detailed log - Double click to close this message]
Failed to load model class 'HBoxModel' from module '@jupyter-widgets/controls'
Error: Module @jupyter-widgets/controls, version ^1.5.0 is not registered, however, 2.0.0 is

my conda version is:
conda version : 4.14.0
python version : 3.9.12.final.0

have I done something to my PC that is preventing me from even recreating my local conda env? what should I wipe to at least get the generate button back? I am at a loss. I have deleted my env folders. should I delete my conda package cache too?

Hello, I have followed the instruction to download the program in this thread. But when I try to run it, I can't get any further than for the program to be stuck in the running mode, see my screenshot. I have waited for around one hour for the program to finish. Have I missed something or should it take time for the program to start?

Update. Now I have the error message, "fatal: destination path 'stablediffusion-infinity' already exists and is not an empty directory."