Lab 10 BIA

What is tokenization?

• A way of separating a piece of text into smaller units.

• A way of changing a text in order for it to be readable.

• A method of separating a word into letters.

• A way of separating a word into silables.

What is the goal of tokenization?

• Creating a vocabulary

• Reducing the dataset

• Generate an RNN suitable dataset

• Create a connection between the tokens and the relation between them

Chose the TRUE statements

• The RNN is a stateful neural network

• The Recurent Neural Network is the same as the Recursive Neural Network

• RNN can be used for handwriting recognition or speech recognition

• For RNNs their neuron is said to have connections between passes, and through time

What is this?

• RNN

• GRU

• NLP

• LSTM

What is this?

• RNN

• GRU

• NLP

• LSTM

What is this?

• RNN

• GRU

• NLP

• LSTM

Lets say you have a vocalubary variable vocab.

Write a code in order to:

- have 512 hidden layers

- 1000 epochs with a learning rate of 0.5

- train using an LSTM layer on a RNN model