https://github.com/arkasarkar19/neural-machine-translation-with-attention-model

Refer Readme.md
https://github.com/arkasarkar19/neural-machine-translation-with-attention-model

attention-mechanism bi-lstm-model loop lstm-neural-networks machine-learning natural-language-processing neural-machine-translation neural-network python rnn-tensorflow step-attention

Last synced: 2 months ago
JSON representation

Refer Readme.md

• Host: GitHub
• URL: https://github.com/arkasarkar19/neural-machine-translation-with-attention-model
• Owner: ArkaSarkar19
• Created: 2020-06-17T20:46:48.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2020-06-18T16:11:32.000Z (almost 5 years ago)
• Last Synced: 2025-01-19T07:15:13.558Z (4 months ago)
• Topics: attention-mechanism, bi-lstm-model, loop, lstm-neural-networks, machine-learning, natural-language-processing, neural-machine-translation, neural-network, python, rnn-tensorflow, step-attention
• Language: Python
• Homepage:
• Size: 249 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Neural-Machine-Translation-with-Attention-Model

The Neural Machine Translation (NMT) model translates human-readable dates ("25th of Feb, 1999") into machine-readable dates ("1999-02-25") 

The code is in Python 3. 

### Model Overview and Implementation Details

**model.py** has 2 main functions : one_step_attention() and model(). 

* one_step_attention() computes :  

    * [α,α,...,α]the attention weights 

    * context^⟨t⟩: the context vector 

* model() function 

   * Model first runs the input through a Bi-LSTM to get [a<1>,a<2>,...,a][a<1>,a<2>,...,a].

   * Then, model calls one_step_attention() Ty times using a for loop. At each iteration of this loop:

       * It gives the computed context vector contextcontext to the post-attention LSTM.

       * It runs the output of the post-attention LSTM through a dense layer with softmax activation.

       * The softmax generates a prediction ŷ 

  

  

![ntm](readme_images/attn_model.png)

### How to Run and Testing

**model.py** has the main code and just run it on the command line. 

In model.py I have added some examples as a test set. You can modify that to test your own examples. 

* Sample Output 

   * After training for 20 epochs. 

   

            source: 3 May 1979

            output: 1979-05-03

            source: 5 April 09

            output: 2009-04-05

            source: 21th of August 2016

            output: 2016-08-01

            source: Tue 10 Jul 2007

            output: 2007-07-10

            source: Saturday May 9 2018

            output: 2018-05-09

            source: March 3 2001

            output: 2001-03-03

            source: March 3rd 2001

            output: 2001-03-01

            source: 1 March 2001

            output: 2001-03-01

### Note

The code might take a couple of minutes to run as it trains 20 epochs.