This library provides smart emoji recommendations on the input text using novel Natural Language Processing methods.
Built by Neetu Kumari
The dataset consists of a list of emojis & their names/descriptions.
This project use BERT which is a state-of-the-art language model for NLP. Emoji's description is tokenized using BERT & stored in a file.
The input text is first normalized by removing the stopwords. The mean of the tokenized values is then taken for the remaining text. The mean tokenized vector of the text is then compared with the tokenized value of each emoji. Using cosine similarity we then find the top five nearest valued emoji.
- Download the repository
- Run the following command
streamlit run streamlit-emoji-recommender.py
- PyTorch
- Hugging Face Transformers
- BERT
- NLTK
- Pandas
- NumPy
If you have any questions or feedback, feel free to reach out to me at [email protected]