Translator

English-Chinese Translator with LSTM and Seq-to-Seq Models

Welcome to my project! 🎉 Here, I've built a translator that converts English to Traditional Chinese and vice versa using advanced deep learning models. It’s an exciting dive into the world of natural language processing (NLP) and machine translation.

---

Project Overview

This project includes:

• Seq-to-Seq Model:
  A sequence-to-sequence model with encoder-decoder architecture for language translation.

• LSTM Model:
  An LSTM-based translation model designed to handle sequential data and long-term dependencies.

• Evaluation Metrics:

  • BLEU Score: Measures the quality of machine translation by comparing model output with reference translations.
  • ChrF Score: Focuses on character-level similarity, particularly useful for non-alphabetic languages like Chinese.

• Dataset:
  A bilingual dataset of 1000+ English-Traditional Chinese sentence pairs, split for training, validation, and testing.

---

How It Works

1. Data Preparation:

   • Sentences are tokenized and padded to maintain uniformity.
   • The dataset is split into training and validation sets.

2. Model Training:

   • Two models are trained:
     • LSTM Model for baseline performance.
     • Seq-to-Seq Model with an attention mechanism for enhanced results.

3. Translation Process:

   • Input a sentence in English or Chinese.
   • The model generates a translation in the target language.

4. Evaluation:

   • Use BLEU and ChrF scores to validate model performance.
   • Plot training and validation loss curves to monitor learning.

---

What I Learned

• Designing and implementing Seq-to-Seq and LSTM models for translation.
• Working with bilingual datasets and tokenizing for both English and Chinese.
• Understanding and applying BLEU and ChrF scores to evaluate translation quality.
• Managing challenges of long sequences and context switching in language models.

---

Results

• BLEU Score: [Add the result here]
• ChrF Score: [Add the result here]

The scores show promising results, with potential for further optimization.

---

Future Work

1. Experimenting with Transformers:
   Implementing models like BERT or GPT to enhance translation quality.

2. Expanding Dataset:
   Adding more sentence pairs to improve fluency and context handling.

3. Multi-language Translation:
   Extending support to other languages like Spanish or French.

---

Usage

1. Clone this repository:

   git clone <repository-link>
   

2. Install required libraries:

   pip install -r requirements.txt
   

3. Run the training scripts:

   python Seq_to_seq_code.ipynb
   python LSTM_code.ipynb
   

4. Input your test sentences and generate translations.

---

Thanks for exploring my project! 🌟 Feel free to fork the repo, try it out, and share your feedback. 😊