https://github.com/junteudjio/amazon_reviews_summarizer
A deep learning based text summarizer for Amazon reviews in tensorflow.
https://github.com/junteudjio/amazon_reviews_summarizer
deep-learning seq2seq tensorflow text-summarization
Last synced: 12 months ago
JSON representation
A deep learning based text summarizer for Amazon reviews in tensorflow.
- Host: GitHub
- URL: https://github.com/junteudjio/amazon_reviews_summarizer
- Owner: junteudjio
- License: mit
- Created: 2018-01-05T03:04:15.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2018-01-05T03:07:21.000Z (over 8 years ago)
- Last Synced: 2025-03-28T17:57:26.899Z (over 1 year ago)
- Topics: deep-learning, seq2seq, tensorflow, text-summarization
- Language: Python
- Size: 46.9 KB
- Stars: 3
- Watchers: 2
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Amazon Reviews Summarizer (ars)
A deep learning based text summarizer for Amazon reviews in tensorflow.
## Synopsis
- In this project, we use the amazon fine food reviews dataset to implement a text summarizers for these reviews.
- The code is meant to be high quality, clean and flexible enough to handle testing different kinds of architecture easily.
- To achieve this goal we use the same coding style as in the beautiful official tensorflow's example for neural machine translation,
from which we have used/adapted several code snippets.
## Architecture overview
- A seq2seq model with bi-directional, multi layers RNN/GRU/LSTM cells.
- Attention mechanism on the decoder.
- 3 different graphs for train, evaluation and test modes (more work but makes code clean and fast).
- used the beautiful dataset/iterator for input data feeding.
- used glove vectors for embeddings initialization.
## Requirements
Install all needed dependnecies through
`pip install -r requirements.txt`.
Or
`python setup.py develop`.
## Running the code
- You can get started by downloading the datasets and doing some basic preprocessing:
$ code/get_started.sh
Note that you will always want to run your code from the "ars" directory, not the code directory, like so:
$ python code/train.py
This ensures that any files created in the process don't pollute the code directoy.
- Now train/evaluate/test the model by running :
$ python code/run_ars.py
change the cmd line args to try different architecture flavours.
## Contributors
- junior Teudjio Mbativou : https://www.linkedin.com/in/junior-teudjio-3a125b8a
# BibTex and Acknowledgment
```
@article{luong17,
author = {Minh{-}Thang Luong and Eugene Brevdo and Rui Zhao},
title = {Neural Machine Translation (seq2seq) Tutorial},
journal = {https://github.com/tensorflow/nmt},
year = {2017},
}
```
## License
This project is licensed under the MIT License - see the [LICENSE.md](LICENSE.md) file for details