🤗 HuggingFace Ecosystem - NLP Playground

A Streamlit-based web application showcasing various NLP capabilities using HuggingFace's Transformers library.

Features

• Sentiment Analysis: Analyze the sentiment of text (positive/negative)
• Text Summarization: Generate concise summaries of longer texts
• Named Entity Recognition: Identify entities like persons, organizations, and locations
• Question Answering: Extract answers to questions from a given context
• Text Generation: Create coherent text continuations from a prompt
• Semantic Search: Find semantically similar texts using embeddings

Models Used

• Sentiment Analysis: distilbert-base-uncased-finetuned-sst-2-english
• Text Summarization: facebook/bart-large-cnn
• Named Entity Recognition: dslim/bert-base-NER
• Question Answering: deepset/roberta-base-squad2
• Text Generation: gpt2
• Semantic Search: sentence-transformers/all-MiniLM-L6-v2

Demo Snapshots

Note: Installation and setup instructions to run this project locally can be found here

Deployed on HuggingFace Spaces

This app is currently deployed as a HuggingFace Space, you can view it here! 🚀

 

Message from the author

This app is something I've always wanted to make ever since I came across the HF ecosystem 2 years ago, especially the different NLP capabilities that it offers (now there are more, but i wanted to start with this). I hope to introduce more features into this as I learn more in the future. So yeah, I am super proud of how this turned out. Although I'm sure there might be more similar apps such as this. But its been on my mind for a long time and I'm glad its been built now. There's ofcourse room for improvement, I look forward to change and experiment things as I move forward with this.