Passing batches to PyTorch Model

Question

I am trying to train a PyTorch LSTM model which is defined as:

class RecurrentNet(torch.nn.Module):
    def __init__(self, d_in, d_hidden, sequence_length):
        super(RecurrentNet, self).__init__()
        self.d_in = d_in
        self.d_hidden = d_hidden
        self.sequence_length = sequence_length
        self.lstm = torch.nn.LSTM(input_size=d_in,  hidden_size=d_hidden, num_layers=sequence_length)
        self.fc = nn.Linear(d_hidden, 1)
        
    def forward(self, x):
        out, hidden =  self.lstm(x)
        y_pred = F.relu(self.fc(out[-1][-1]))
        return y_pred

model = RecurrentNet(28, 3, l)

I have prepared my data in the form of batches of len 32, and with each instance having dimensions of 6 x 28, i.e. torch.Size([6, 28]). This means the total input tensor is size torch.Size([32, 6, 28]), with labels size torch.Size([32, 1]).

When I pass a single instance to the untrained model it returns an integer as expected. When I pass the Tensor of 32 intances, it returns:

tensor([0.], grad_fn=)

I would have expect a 32 x 1 output Tensor for this input. Have I missed something when preparing the training data?

Passing batches to PyTorch Model

0 Answers