Move window tensor to proper device

[hhatefi]

When running inference on GPU, Wav2Win._forward_strided fails if x_strided and self._window are on different device. Therefore, self._window should move into the same device as that of x_strided

[pzelasko]

You can move the Wav2Win (or any other module in that file) to device before running the inference. If you're using Fbank, there's an option in the config to place it on the gpu.

[hhatefi]

The problem it not the module, it is self._window which is created by create_frame_window which does not set the device type. Even if the module is loaded into gpu, forward call will fail, if self._window is not in gpu.

[pzelasko]

Wouldn't module = module.to("cuda") solve the issue by placing every registered parameter and buffer on the GPU?