Installation for window

[codetolearner]

It seems that the "conda create --name look2hear --file look2hear.yml" the dependencies are only for linux. Is there a way to install that or a guide that show us how to install it in window and instead of using conda. If it is possible to install the dependencies via pip?

[YCbCr444]

I have same issue as you, I tried WSL on Windows (Ubuntu 24.04.2 LTS) and it still doesn't work for me.

The error:

(base) user@DESKTOP-8I5TJ8U:~/Apollo$ conda create --name look2hear --file look2hear.yml
Channels:
 - defaults
Platform: linux-64
Collecting package metadata (repodata.json): done
Solving environment: failed

PackagesNotFoundError: The following packages are not available from current channels:

  - aiohappyeyeballs==2.4.0=pypi_0

Current channels:

  - https://repo.anaconda.com/pkgs/main
  - https://repo.anaconda.com/pkgs/r

To search for alternate channels that may provide the conda package you're
looking for, navigate to

    https://anaconda.org

and use the search bar at the top of the page.

[LeiZhu666]

same question

[JusperLee]

Install PIP Packages Manually After Conda Creates the Base Environment

conda create -n look2hear python=3.11
conda activate look2hear
pip install aiohappyeyeballs==2.4.0

[teodly]

I created Dockerfile & helper scripts. I don't use Windows myself but I've heard that Docker works with WSL. However I'm not sure about GPU support.

https://github.com/teodly/Apollo-Docker

It is for AMD GPU, for Nvidia change base image (FROM) and repo (--index-url). For CPU, you can use my Dockerfile but it will waste disk space for ROCm, I suggest to use plain Python base image and CPU-only repo (https://download.pytorch.org/whl/nightly/cpu/). However, I haven't tried running Apollo on CPU and don't know how slow it will be.

[starinspace]

@codetolearner @YCbCr444 @LeiZhu666

I created an installation tutorial for Windows with a few minor adjustments. Hope this helps!

Apollo-Windows