How to reshape tensors for LSTM input

1

Hi, I have the data like below,
txt_features = np.array([[1,2,3], [4,5,6]])
num_features = np.array([[1,2,4], [4,5,7]])
I want to pass it into a lstm. So how do I reshape the data to c, which is shown below, so that I can feed it into a lstm
c = [
[[1,2,3], [1,2,4]],
[[4,5,6], [4,5,7]]
]

2

You could use torch.stack as seen here:

txt_features = torch.tensor([[1,2,3], [4,5,6]])
num_features = torch.tensor([[1,2,4], [4,5,7]])
c = torch.stack((txt_features, num_features), dim=1)