The line ula, (h_out, _) = self.lstm(x, (h_0, c_0)) causes an exception to be thrown on my machine, i replaced that line with ula, (h_out, _) = self.lstm(x.view(len(x), self.seq_length, -1), (h_0, c_0)). The dimension of the input x should be 3 dimensions while the first line provides only two, the latter line actually modifies the dimension.

Inside LSTM class:

change this:

ula, (h_out, _) = self.lstm(x, (h_0, c_0))
h_out = h_out.view(-1, self.hidden_size)

to this:

h_out, _ = self.lstm(x, (h_0, c_0))
h_out = h_out.view(-1,x.size(1),self.hidden_size)[-1] 

also, when you train/predict the model, use dataX.transpose(0,1) as input

Check https://pytorch.org/docs/stable/nn.html?highlight=lstm#torch.nn.LSTM for reference on input/output formatting for LSTM layer

Now you can use multiple layers.

I think there is something missing. İf I understand correctly implementation of forward is wrong.
At code of forward of LSTM model, we see h_0=torch.zeros(...) and c_0 = torch.zeros(...) which means you set state everytime same, not get any information of sequence state. İs there anything ı am missing?