File size: 1,021 Bytes

a9d908e
bf56260
 
370b74d
 
 
 
 
 
38c223c
9cb7b74
370b74d
 
 
 
 
1f7d8b8
370b74d
 
433ebf3
370b74d
 
cb56a86
 
38c223c
cb56a86
38c223c
cb56a86
5c6f564

---
license: apache-2.0
---
# Text-Summarizer
## About

An Abstractive text summarizer trained using lstm based sequence to sequence model with attention mechanisim. The attention model is used for generating each word of the summary conditioned on the input sentence.

- Used CNN_DailyMail dataset.
- Code + Deployment : https://www.youtube.com/watch?v=LFZBA99NOpU
[![IMAGE ALT TEXT HERE](https://i.ytimg.com/vi/LFZBA99NOpU/hqdefault.jpg?sqp=-oaymwEiCKgBEF5IWvKriqkDFQgBFQAAAAAYASUAAMhCPQCAokN4AQ==&rs=AOn4CLCQacvrk4iyXIEhzufcyLKd9ZWHlQ)](https://www.youtube.com/watch?v=LFZBA99NOpU)

### Training Model Overview

loss graph

![Output](./train_log.jpeg "loss graph")

encoder-decoder overview
![Output1](./model_plot.jpeg "model overview")

## Conclusion

- 🫶  The machine learning model to convert a text document to abstract is done successfully.
- 🫶  Created a Flask app using an api call from this repository & deployed the app in heroku app.

## Deployment:

- 🫶  https://text-summariser-v1.herokuapp.com/