Batched Interleaved Model Using N images to train the ConvLSTM, we train this model over the lane segementation dataset. This is our model: This code is written in lightening and PyTorch.