Encoder-Decoder NLP Suite

NMIMS MPSTME · Advanced ML · Experiment 6 · Dr Ami Munshi

Task	Description	Model
T1 & T2	English → Hindi + BLEU eval	Custom LSTM Seq2Seq
T3	English → Spanish	MarianMT (Helsinki-NLP)
T4	Text Summarization	BART (facebook/bart-large-cnn)

---

Setup & Run

# 1. Install
pip install flask flask-cors torch transformers datasets sacrebleu sentencepiece

# 2. (Optional) Train En→Hi model — needs data/english_hindi.csv from Kaggle
#    https://www.kaggle.com/code/uselessnoob/english-to-hindi-machine-translation
python app.py --train-hi

# 3. Start backend
python app.py          # → http://localhost:5000

# 4. Open frontend
open index.html        # or just double-click it

---

Push to GitHub (new repo)

# Requires GitHub CLI: https://cli.github.com
git init && git add -A
git commit -m "feat: Encoder-Decoder NLP Suite — Experiment 6"
gh repo create encoder-decoder-nlp --public --source=. --remote=origin --push

---

API Endpoints

Method	Endpoint	Body
GET	/api/health	—
POST	/api/translate/en-hi	{ "text": "..." }
POST	/api/translate/en-es	{ "text": "..." }
POST	/api/summarize	{ "text": "...", "max_length": 150 }