ADARNN torch CPU only

[kurucan]

How can I set to use CPU only If GPU is not available? The error stems from the line 150 in pytorch_adarnn.py

/usr/local/lib/python3.7/site-packages/pyqlib-0.7.2.99-py3.7-linux-x86_64.egg/qlib/contrib/model/pytorch_adarnn.py in init(self, d_feat, hidden_size, num_layers, dropout, n_epochs, pre_epoch, dw, loss_type, len_seq, len_win, lr, metric, batch_size, early_stop, loss, optimizer, n_splits, GPU, seed, **kwargs) 148 149 self.fitted = False --> 150 self.model.cuda() 151 152 @property

RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from

https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/pytorch_adarnn.py self.device = torch.device("cuda:%d" % (GPU) if torch.cuda.is_available() and GPU >= 0 else "cpu")

[you-n-g]

Thanks for reporting the bug :) I think replacing all the .cuda() with .to(self.device) will solve this issue. You can follow this example. https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/pytorch_nn.py#L157

Would you mind sending a PR to fix this issue and become one of the contributors of Qlib!

Thanks

[github-actions[bot]]

This issue is stale because it has been open for three months with no activity. Remove the stale label or comment on the issue otherwise this will be closed in 5 days

[you-n-g]

@SunsetWolf will help to solve this issue

[HyeongminMoon]

I think this issue already fixed&merged in #1070 . Can I ask why @you-n-g reopened this?

[github-actions[bot]]

This issue is stale because it has been open for three months with no activity. Remove the stale label or comment on the issue otherwise this will be closed in 5 days