ksm26/function-calling-and-data-extraction-with-llms

Master the techniques of function-calling and structured data extraction with LLMs. Learn to enhance LLM capabilities, integrate web services, and build practical applications for real-world data usability.

📊 Function-calling and Data Extraction with LLMs

💡 Welcome to the "Function-calling and Data Extraction with LLMs" course! The course will equip you with the critical skills for building advanced applications with LLMs.

Course Summary

In this course, you'll dive into the essentials of function-calling and structured data extraction with LLMs, focusing on practical applications and advanced workflows. Here's what you can expect to learn and experience:

1. 🛠️ Function-calling: Learn to extend LLMs with custom capabilities by enabling them to call external functions based on natural language instructions, using NexusRavenV2-13B, an open-source model fine-tuned for function-calling and data extraction.

2. 🔄 Complex Workflows: Work with multiple function calls, including parallel and nested calls, to create complex agent workflows where an LLM plans and executes a series of functions to achieve a goal.
3. 🌐 Web Services Integration: Use OpenAPI specifications to build function calls that can access web services, enhancing the functionality and reach of your applications.
4. 🗂️ Structured Data Extraction: Extract structured data from natural language inputs, enabling real-world data usability for analysis and application.

5. 💾 End-to-End Application: Build an application that processes customer service transcripts, generates SQL calls, and stores results in a database, demonstrating the practical implementation of the skills learned.

Key Points

• 🔌 Extend LLM Functionality: Learn to extend LLMs with custom functionality via function-calling, enabling them to perform external function calls.
• 📊 Data Usability: Extract structured data from natural language inputs, making real-world data usable for analysis.
• 🛠️ Practical Implementation: Build an end-to-end application that processes customer service transcripts using LLMs.

About the Instructors

🌟 Jiantao Jiao is the Co-founder & CEO of Nexusflow and an Assistant Professor of EECS and Statistics at UC Berkeley, bringing extensive expertise in function-calling and data extraction.

🌟 Venkat Srinivasan is a Founding Engineer at Nexusflow, specializing in the development of advanced LLM applications.

🔗 To enroll in the course or for further information, visit deeplearning.ai.