Processed a 70k+ image crop disease dataset, utilized transfer learning and finetuned the mobilenet (5 million parameters) for disease recognition. Developed a streamlit web platform for the model, integrated with GPT-4 for farmer interactions and chatting about remedies
View on GitHub
Developed within 36 hours, we finetuned a ViT-GPT 2 (200 million parameters) using a food images and ingredients dataset, integrating with GPT-4 API for nutrition insights. Rolled out a user friendly interface using StreamLit, hosting the model on Hugging Face.
View on GitHub
Developed a dynamic platform for food exploration, using React for a responsive UI and JavaScript for advanced search functionalities, aiding in recipe exploration and dietary filtering. Leveraged HTML/CSS to design the layout, ensuring an engaging user experience.
View on GitHub
Developed a web tool with Flask to dynamically generate text summaries, utilizing SpaCy and NLTK for natural language processing. Enhanced user interactivity with JavaScript, offering precise keyword extraction for quick content insights.
View on GitHub