https://github.com/skrishna-7/nextword-predictor
Simple Next Word predictor with LSTM
https://github.com/skrishna-7/nextword-predictor
deep-learning deeplearning-projects lstm project streamlit
Last synced: 3 months ago
JSON representation
Simple Next Word predictor with LSTM
- Host: GitHub
- URL: https://github.com/skrishna-7/nextword-predictor
- Owner: SKrishna-7
- Created: 2025-02-15T14:20:17.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2025-02-15T14:36:41.000Z (3 months ago)
- Last Synced: 2025-02-15T15:31:52.119Z (3 months ago)
- Topics: deep-learning, deeplearning-projects, lstm, project, streamlit
- Language: Jupyter Notebook
- Homepage: https://skrishna-7-nextword-predictor-app-mk9wgi.streamlit.app/
- Size: 12.7 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Next-Word Predictor using LSTM
## Overview
This project implements a next-word prediction model using Long Short-Term Memory (LSTM) networks. The dataset used for training is *The Tragedie of Hamlet by William Shakespeare (1599)*. The model takes a sequence of words as input and predicts the most probable next word.
## Dataset
The dataset consists of the text from *Hamlet*, preprocessed to remove unnecessary characters and formatted into sequences for training the LSTM model.
## Technologies Used
- **Python**
- **TensorFlow & Keras**
- **Natural Language Processing (NLP)**
- **LSTM Neural Networks**
- **Streamlit** (for deployment)
## Workflow
* Data Collection: We use the text of Shakespeare's Hamlet as our dataset. This rich, complex text provides a good challenge for our model.
* Data Preprocessing: The text data is tokenized, converted into sequences, and padded to ensure uniform input lengths. The sequences are then split into training and testing sets.
* Model Building: An LSTM model is constructed with an embedding layer, two LSTM layers, and a dense output layer with a softmax activation function to predict the probability of the next word.
* Model Training: The model is trained using the prepared sequences, with early stopping implemented to prevent overfitting. Early stopping monitors the validation loss and stops training when the loss stops improving.
* Model Evaluation: The model is evaluated using a set of example sentences to test its ability to predict the next word accurately.
* Deployment: A Streamlit web application is developed to allow users to input a sequence of words and get the predicted next word in real-time.