CNN LSTM implementation for video classification

vision
7 
def forward(self, x):

    batch_size, timesteps, C,H, W = x.size()
    c_in = x.view(batch_size * timesteps, C, H, W)
    
    c_out = self.cnn(c_in)

    r_out, (h_n, h_c) = self.rnn(c_out.view(-1,batch_size,c_out.shape[-1]))  

    logits = self.classifier(r_out)

    return logits