This repository contains a Handwritten Mathematical Expression (HME) recognition system that translates handwritten mathematical expressions into sequences of symbols and relations. The project implements a Recurrent Neural Network (RNN) with bidirectional LSTM cells and uses Connectionist Temporal Classification (CTC) loss for training.
- PyTorch Lightning
- Bidirectional LSTM
- Connectionist Temporal Classification (CTC)
- CROHME 2019 dataset
- Weights & Biases for experiment tracking
IMPORTANT NOTICE: This work is submitted as part of a university assignment. It is strictly prohibited to copy, modify, and/or pass this work as your own. This repository is made available for educational and reference purposes only. Any unauthorized use of this code for academic submissions constitutes plagiarism and violates academic integrity policies.