Training the LSTM

Now that the video features are generated, we can use them to train an LSTM. This step is very similar to the training steps described earlier in the book—we define a model and an input pipeline, and launch the training.