informal_rnn.py

rnn = nn.RNNCell(3, 5) # n_input X n_neurons

X_batch = torch.tensor([[[0,1,2], [3,4,5], 
                         [6,7,8], [9,0,1]],
                        [[9,8,7], [0,0,0], 
                         [6,5,4], [3,2,1]]
                       ], dtype = torch.float) # X0 and X1

hx = torch.randn(4, 5) # m X n_neurons
output = []

# for each time step
for i in range(2):
    hx = rnn(X_batch[i], hx)
    output.append(hx)

print(output)

### output
'''
[tensor([[-0.4872,  0.4388, -0.3683, -0.2402, -0.7824],
        [-0.9869, -0.9002, -0.9990, -0.5977, -0.9543],
        [-0.9989, -0.9978, -0.9999, -0.3553, -0.9991],
        [-0.9998, -0.9889, -0.9749, -0.9821, -0.5408]],
       grad_fn=<TanhBackward>), tensor([[-0.9998, -0.9983, -0.9999, -0.3426, -0.9987],
        [ 0.2678,  0.8609, -0.3364, -0.0767,  0.4827],
        [-0.9897, -0.9457, -0.9979, -0.2552, -0.9282],
        [-0.6595, -0.2529, -0.8555, -0.1959, -0.2725]],
       grad_fn=<TanhBackward>)]
'''