Part 2: Let's play a few more games
Great work on the last project. This time, you will need to simulate multiple game shows hosted over several days. This means that your data will have a time dimension to it. More specifically, your data will have the shape (number of contestants, game shows, inputs size
).
You will need to extend your model to incorporate this new feature. For this you will be using a TimeDistributed
layer to allow the Dense
layer to accept contestants from multiple game shows.
You have been provided with the weight initializer init
, the prizes
list from the previous exercise, a time-series input x
and names
which contains the names of the contestants. x
is a (3,2,3)
numpy array where names
is a (2,3)
Python list. In other words, you have 2 game shows (i.e. sequence length), each with 3 contestants (batch size) where each contestant has 3 attributes (input size).
This exercise is part of the course
Machine Translation with Keras
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Print names and x
print('names=\n',____, '\nx=\n',____, '\nx.shape=', ____)