cannot train with multiple gpus

I recently run the code with two RTX A6000 gpus:
I got this error:
RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:1!

This is my related environment:
python 3.8.12 h12debd9_0
pytorch 1.8.0 py3.8_cuda11.1_cudnn8.0.5_0 pytorch
slowfast 1.0 dev_0
torchaudio 0.8.0 py38 pytorch
torchvision 0.9.0 py38_cu111 pytorch

I also print the "torch.cuda.current_device()" in build.py, which always shows "0".

Could you please help me? Thanks!!