Search⌘ K
AI Features
Home
Courses
Building Advanced Deep Learning and NLP Projects

Project Creation: Part Two

Explore how to convert sentences into 50-dimensional vectors using GloVe embeddings and one hot encode outputs to prepare data for an emoji prediction model. Understand the data preprocessing steps and get ready to build an LSTM architecture for NLP project development.

We'll cover the following...

In the previous lesson, we loaded the word vectors into a dictionary. We will now create a function that will accept the data and return a vector representation for each sentence in the data. Before moving on, we also need to convert our dataframe into the train and test the input and output. The output needs to be in one hot encoding.

Create the input and output data

Python 3.5
from tensorflow.python.keras.utils import to_categorical
XT = train[0]
Xt = test[0]
YT = to_categorical(train[1])
Yt = to_categorical(test[1])
print(XT.shape)
print(Xt.shape)
print(YT.shape)
print(Yt.shape)

Explanation:

  • We have used Keras to create one hot encodings to the
...