Possible bug in data processing generating incorrect optical flow

[lzhyu]

ReCoNet-PyTorch/totaldata.py

Lines 112 to 115 in 8d6d98a

	originalflow=torch.from_numpy(flow)
	flow = torch.from_numpy(transform.resize(flow, (360, 640))).to(device).permute(2,0,1).float()
	flow[0, :, :] *= float(flow.shape[1])/originalflow.shape[1]
	flow[1, :, :] *= float(flow.shape[2])/originalflow.shape[2]

Here the shape of flow has been permuted, so it should be
flow[0, :, :] *= float(flow.shape[1])/originalflow.shape[0],
flow[1, :, :] *= float(flow.shape[2])/originalflow.shape[1]