[Feature Request]: Is it any way to run it on Windows natively?

[ke1ne]

Is there an existing issue for this?

• I have searched the existing issues and checked the recent builds/commits

What would your feature do ?

I would like to run your fork on Windows without WSL. Is it possible, basically?
Thank you!

Proposed workflow

1. Run some cmd-file.
2. Get WebUI, generation, etc. working.

Additional information

No response

[vladmandic]

i've just re-added original webui.bat

[ecker00]

The readme is missing documentation on how to install in windows.

Edit: If I understand correctly it's to just run webui.bat? But that gives me:

venv "…\automatic\venv\Scripts\Python.exe"
Traceback (most recent call last):
  File "…\automatic\launch.py", line 12, in <module>
    from rich import print
ModuleNotFoundError: No module named 'rich'

Which sounds a bit like python dependencies are not installed. Am I missing an instruction step?

[ke1ne]

@ecker00

1. Create&activate virtual environment:

cd automatic
python -m venv venv
.\venv\Scripts\activate

2. Install Torch into that venv:

pip install torch torchaudio torchvision triton --force --extra-index-url https://download.pytorch.org/whl/cu118

3. Exit venv:

.\venv\Scripts\deactivate

3. Run UI via webui.bat. I'm using own runner instead (like start.bat):

@echo off

call git pull --recurse-submodules

set TF_CPP_MIN_LOG_LEVEL=2
set FORCE_CUDA="1"
set ATTN_PRECISION=fp16
set PYTORCH_CUDA_ALLOC_CONF=garbage_collection_threshold:0.9,max_split_size_mb:512
set CUDA_LAUNCH_BLOCKING=0
set CUDA_CACHE_DISABLE=0
set CUDA_AUTO_BOOST=1
set CUDA_DEVICE_DEFAULT_PERSISTING_L2_CACHE_PERCENTAGE_LIMIT=0
set CUDA_MODULE_LOADING="LAZY"
set GRADIO_ANALYTICS_ENABLED="False"
set SAFETENSORS_FAST_GPU=1

set COMMANDLINE_ARGS=--ckpt-dir="D:\AI\models" --vae-dir="D:\AI\vaes" --autolaunch --opt-channelslast --upcast-sampling --api --gradio-queue --theme dark --cors-allow-origins="http://127.0.0.1:7860" --update-all-extensions

call webui.bat

[ecker00]

That seemed to almost do it, but get error that it's unable to find the required version:

pip install torch torchaudio torchvision triton --force --extra-index-url https://download.pytorch.org/whl/cu118

Collecting torch
  Downloading torch-2.0.0-cp310-cp310-win_amd64.whl (172.3 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 172.3/172.3 MB 12.8 MB/s eta 0:00:00
Collecting torchaudio
  Downloading torchaudio-2.0.1-cp310-cp310-win_amd64.whl (2.1 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.1/2.1 MB 16.9 MB/s eta 0:00:00
Collecting torchvision
  Downloading torchvision-0.15.1-cp310-cp310-win_amd64.whl (1.2 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.2/1.2 MB 9.5 MB/s eta 0:00:00
ERROR: Could not find a version that satisfies the requirement triton (from versions: none)
ERROR: No matching distribution found for triton

If this works, should really add it to the readme.md

[ke1ne]

That seemed to almost do it, but get error that it's unable to find the required version:

Oh, simply remove the triton from the pip install line. There is no triton lib available on Windows.

[ecker00]

Thank you, that got things to progress a bit!

But now encountering another error, but that's tracked here:
#100

[ke1ne]

@ecker00 It looks like we have to be patient a couple of days, there are multiple changes happening with the project, and many possible issues too. But it's great anyway.