Ingesting a large dataset and preprocessing it efficiently
What is the objective of this command dataset.repeat(3).batch(7)
Answer :
repeat() function repeat all the
elements 3 times and then create
the batch in such a way that each
batch will have 7 elements
A tensor is a multidimensional array that can have more than 2 axes as well.
Answer :
Correct
Due to the transformations that the data goes through when traversing an RNN, some information is lost at each time step.
Answer :
Correct
In which keras API an instance of the Sequential class is created and model layers are created and added to it?
Answer :
Sequential API
Activation functions are mathematical equations that determine the output of a neural network.
Answer :
Correct
What type of RNN is this?
"At each training iteration the model starts with a hidden state full of zeros"
Answer :
Stateless RNNs
In which type of neural network, two neural networks contesting with each other?
Answer :
GAN
Let's say you want to use shuffle on dataset of 10,000 observation and your buffer size is 1000. What will happen?
Answer :
Shuffle will initially select a random element from first 1000 elements in the buffer. Once an element is selected, it’s space in the buffer is replaced by the next (i.e. is 1001st element) maintaining the 1000 element buffer
The main protobuf typically used in a TFRecord file is the Example protobuf, which represents one instance in a dataset
