Smanj/LSTM_Seq2Seq

Build Your Own Simple Translator and Seq-to-Seq Translator

This project provides a guide to building two types of neural machine translation (NMT) models using sequence-to-sequence (Seq-to-Seq) learning. It covers both a foundational simple translator model and a more advanced Seq-to-Seq model with an attention mechanism, ideal for translating text between languages. This repository contains the code and resources for building two types of neural machine translation models using sequence-to-sequence (Seq-to-Seq) learning.

Project Overview This project walks through the steps to create two different machine translation models: Simple Translator: A basic model that demonstrates the core principles of machine translation, using a straightforward encoder-decoder structure. Seq-to-Seq Translator with Attention: An advanced translation model utilizing the attention mechanism to improve translation accuracy by dynamically focusing on relevant parts of the input sequence.

Dataset Preparation

Dataset: Use a parallel text dataset, such as the English-Finnish data. For English https://huggingface.co/datasets/yahma/alpaca-cleaned and for Finnish https://huggingface.co/datasets/saillab/taco-datasets/tree/main/multilingual-instruction-tuning-dataset%20/multilingual-alpaca-52k-gpt-4/Finnish was used Preprocess the Data:

1. Tokenize and clean text data.
2. Prepare pairs in a tab-separated format, where each line contains a source and target sentence pair.

How to Use

1. Clone the repository:

   git clone <repository_url>
   cd <repository_directory>
   Install the required packages: Ensure you have the necessary libraries installed. Can be done by running:
   

pip install -r requirements.txt Run the Jupyter Notebook: Open the notebook in Jupyter:

jupyter notebook LSTM_based_translator.ipynb and Seq2Seq_based_translator.ipynb