Skip to content

hpanwar08/sentiment-analysis-torchtext

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Sentiment Analysis in Torch Text

Sentiment analysis is a classification task where each sample is assigned a positive or negative label.

This repo contains the code for the this blog.

Typical components of classification task in NLP

  1. Preprocessing and tokenization
  2. Generating vocabulary of unique tokens and converting words to indices
  3. Loading pretrained vectors e.g. Glove, Word2vec, Fasttext
  4. Padding text with zeros in case of variable lengths
  5. Dataloading and batching
  6. Model creation and training

Why use torchtext

Torchtext provide set of classes that are useful in NLP tasks. These classes takes care of first 5 points above with very minimal code.

Prerequisites

What is covered in the notebook

  1. Train validation split
  2. Define how to process data
  3. Create torchtext dataset
  4. Load pretrained word vectors and building vocabulary
  5. Loading the data in batches
  6. Simple GRU model
  7. GRU model with concat pooling
  8. Training

Data Overview

Top 5 rows of dataset

Concat Pooling model architecture [2]

GRU model with concat pooling

References

[1] https://colah.github.io/posts/2015-08-Understanding-LSTMs/
[2] https://arxiv.org/abs/1801.06146

Releases

No releases published

Packages

No packages published