Developed a text-to-SQL query generator and conversational chatbot using the Mixtral-7B Model. Pre-processed and tokenized input text data to ensure compatibility, leading to a 40% improvement in SQL query generation accuracy. Implemented Retrieval-Augmented Generation (RAG) by creating a specialized database for enhanced query generation and database interaction, which reduced average processing times by 10%.
The project aimed to bridge the gap between natural language queries and SQL, making it easier for users to interact with databases using conversational language. By leveraging state-of-the-art language models and advanced preprocessing techniques, the solution achieved significant improvements in query accuracy and efficiency.