Closed olegchomp closed 2 years ago
Hello @4omp. The --cuda
argument is not in the list of available console interface arguments. As you can see in our README.md and in the CLI help message, the --device cuda
flag can be used to start processing in gpu mode.
Upd. In any other case, if the argument in your post is typo, I tested it now on real hardware with docker and everything works as expected. Check your Python environment setup.
Tried to run on a local PC with anaconda and CLI --cuda, but seems no GPU utilization. PyTorch recognize card and it works in others scripts.