Cuda unavailable issue when running the demo.py with parameters of 'cuda device'

[FlemingH]

[FlemingH]

1. Change default parameters of --device to “cuda device”

2. Run the demo.py

[FlemingH]

If you print the 'torch.cuda.is_available()' before select_device function

Problem Solved

[FlemingH]

可以检查一下为什么嘛，谢谢！

[akashchgupta]

If you print the 'torch.cuda.is_available()' before select_device function

Problem Solved

And, why exactly is that?

[FlemingH]

@akashchgupta If you're having the same output, that's what I'm asking

[danielzhangau]

1. Change default parameters of --device to “cuda device”
2. Run the demo.py

the 'Cuda available' is the description, you should choose from 0,1,2 or CPU. If you have only one GPU, then just use 0 as the default option.

[husaixy]

A new possibility is that if your environment is anaconda, the default torch installation is the cpu version

[KNzzzzzzzzzz]

This can solve the problem.