Developed Corpus2GPT, a pioneering NLP framework compatible with Keras, TensorFlow, or JAX backends. Enhanced model accessibility and usability with built-in support for multiple languages and Sentence Piece Tokenizer.
View ProjectPython-based implementation of the Transformer architecture from scratch, featuring a bi-directional design similar to BERT and GPT. Used for various NLP tasks including Neural Machine Translation.
View ProjectImplemented Bahdanau-Attention based Seq2Seq NLP Model using TensorFlow and GRU cells for various language translation tasks. Achieved state-of-the-art results in text translation.
View Project