running on windows, how to

[5shekel]

this worked for me. on windows 11 using virtual env (venv).

note the use of py.exe to set the correct python version.
this is similar to the conda create -n gen python=3.10.11 if you want to use conda you can skip this step

you will have to install python-3.10.11-amd64.exe on your windows machine.

# use `py.exe --list` to verify which python version re available to you
py --list
# sample output
 -V:3.11          Python 3.11 (64-bit)
 -V:3.10          Python 3.10 (64-bit)
 -V:3.9
 -V:3.8           Python 3.8 (64-bit)

edit the requirements\pt2.txt to lock some versions.

1. remove the triton package mention and replace with

https://huggingface.co/r4ziel/xformers_pre_built/resolve/main/triton-2.0.0-cp310-cp310-win_amd64.whl

2. set the index-url for xformers, just add this line before the call from xformers. the txt fil should end with some like

--extra-index-url https://download.pytorch.org/whl/cu118
xformers>=0.0.20

then run (in powershell)

py -3.10 -m venv venv  
.\venv\Scripts\activate
pip install -r .\requirements\pt2.txt  
$env:PYTHONPATH = "."
streamlit run scripts/demo/sampling.py --server.port 1234

note: to get the models, d/l from hugginface and place in a new dir named checkpoints, inside the repo.

i use a symlink from another location on my pc so i dont have to have duplicates.

thanks for information in the open/closed issues for some of these bits. specifically https://github.com/Stability-AI/generative-models/issues/159

[cibernicola]

Finally! Thanks, tested and at least it loads! (more tests soon) Windows 10.

[WyattSanders]

https://github.com/Stability-AI/generative-models/issues/172#issue-2007613291 Finally got it working with this running on Windows 11, Conda python 3.10 and then hit this missing scripts error then could load a model: https://github.com/Stability-AI/generative-models/issues/93

[MotorCityCobra]

note: to get the models, d/l from hugginface and place in a new dir named checkpoints, inside the repo.

I tried this.

I'm getting (tf) C:\Users\ooo\tor\generative-models>streamlit run simple_video_sample.py

You can now view your Streamlit app in your browser.

Local URL: http://localhost:8501 Network URL: http://192.168.0.244:8501

C:\ProgramData\Miniconda3\envs\tf\lib\site-packages\torchaudio\backend\utils.py:66: UserWarning: No audio backend is available. warnings.warn("No audio backend is available.") no module 'xformers'. Processing without... No SDP backend available, likely because you are running in pytorch versions < 2.0. In fact, you are using PyTorch 1.11.0+cu113. You might want to consider upgrading. no module 'xformers'. Processing without... 2023-11-26 22:39:43.872 Uncaught app exception Traceback (most recent call last): File "C:\ProgramData\Miniconda3\envs\tf\lib\site-packages\streamlit\runtime\scriptrunner\script_runner.py", line 534, in _run_script exec(code, module.dict) File "C:\Users\ooo\tor\generative-models\simple_video_sample.py", line 278, in Fire(sample) File "C:\ProgramData\Miniconda3\envs\tf\lib\site-packages\fire\core.py", line 141, in Fire component_trace = _Fire(component, args, parsed_flag_args, context, name) File "C:\ProgramData\Miniconda3\envs\tf\lib\site-packages\fire\core.py", line 475, in _Fire component, remaining_args = _CallAndUpdateTrace( File "C:\ProgramData\Miniconda3\envs\tf\lib\site-packages\fire\core.py", line 691, in _CallAndUpdateTrace component = fn(*varargs, **kwargs) File "C:\Users\ooo\tor\generative-models\simple_video_sample.py", line 67, in sample model, filter = load_model( File "C:\Users\ooo\tor\generative-models\simple_video_sample.py", line 268, in load_model with torch.device(device): AttributeError: enter

I get this error before and after the weights are in that directory.

[5shekel]

@MotorCityCobra you don't have xformers. Edit your requirements file. See step 2