Auto-Correct

Overview

AutoCorrect system based upon seq2seq model that is used to correct spelling of wrong words

About Seq2Seq Model

Sequential2Sequential or seq2seq is machine learning model takes sequential input and gives sequential output It is based upon Encoder-Decoder Architecture It is used for phrase-based statistical machine translation (SMT)

Why it is used:

Autocorrection is also need sequence of charaters and need to generate sequence of characters as output which is similar to Machine translation at character level

Key Features

• 1 Deep Learning Model is based upon seq2seq model which is used since autocorrection is comparable to concept of machine translation that is used to convert of language to other
• 2 Natural Language Translation NLP is used in it.

Dependencies:

• Numpy
• Glob
• Tensorflow
• Streamlit

Organisation of files

/project-root
||-- dataset
||-- AutoCorrect.ipynb
||-- Front_end.py
||--Encoder.h5
||-- Decoder.hs
||--README.md

• Dataset : Folder that has dataset of files used to create it
• AutoCorrect.ipynb : Notebook thata consist how the model has been created
• Encoder.h5 and Decoder.h5 : Encoder and Decoder models