PyTorch LSTM -- 2

in order to apply this LSTM API, according pytorch documentation, we need three input parameters, i.e., input_tensor of shape (L,N, H_in) assuming batch first defaulted to false, h_0 of shape (D∗num_layer,N, H_out) and c_0 of shape (D∗num_layers,N,H_cell), where N is batch size, and D= 2 if bidirectional=True otherwise 1. What are the expected value for num_layer and H_in for input_tensor and h_0, respectively?