--gpus=all or equivalent behavior possible with podman-py?

[codeuh]

Hey there podman-py community! I've been working on a personal project that tries to create containers with podman-py. I'm not sure if it's possible to spin containers up with podman-py that mounts in my GPU successfully. I'd love some assistance with this or an answer if it's possible to do what I'm trying to do. My environment is a Red Hat Enterprise Linux WSL2 distro. I'm currently using an NVIDIA RTX 3080. I can run the following command from the WSL2 distro, and the container runs with the GPU as expected.

podman run -d -e "PULSE_SERVER=${PULSE_SERVER}" \
-v /mnt/wslg/:/mnt/wslg/ \
-v /home/codeuh/dev/hue-ai/voice/OpenVoice:/app/src \
-v /home/codeuh/dev/hue-ai/shared/tts/:/app/tts \
--gpus=all \
--name my-voice \
my-voice

I'm running another container where I mount in my WSL2's rootless podman socket for the podman-py client to access it. This other container is a web-ui that I want to manage another container in this project. I'd like to be able to use podman-py to build, create, start, stop, and remove containers from inside this container. I have the logic working, but when I create the podman containers with podman-py, I'm not sure how to get it to create it with GPU access. Here's my function that creates a container in this web-ui.

def create_voice_container(client):
   container = client.containers.create(
       image='my-voice',
       name='my-voice',
       remove=True,
       environment={
           "PULSE_SERVER": os.environ.get("PULSE_SERVER", "")
       },
       mounts=[
           {
               "type": "bind",
               "source": "/mnt/wslg/",
               "target": "/mnt/wslg/"
           },
           {
               "type": "bind",
               "source": "/home/codeuh/dev/hue-ai/voice/OpenVoice",
               "target": "/app/src"
           },
           {
               "type": "bind",
               "source": "/home/codeuh/dev/hue-ai/shared/tts/",
               "target": "/app/tts"
           }
       ]
   )
   return container

Does anyone know if there's a way to do what --gpus=all is doing with the podman-py client in the environment I described?

Thanks for your time and consideration,
codeuh