LLM-Function-Calling-and-Data-Extraction
☰ Table of Contents
- Goal
- Key concepts in the project
- Function calling
- Function calling with external tools
- Structured extraction
- Function calling use cases
- References
Goal
Build a Dialogue Data Extraction System
- Part 1 - Process breakdown
- Defining required data to be extracted
- Building database to store extracted data
- Defining tools to populate the database
- Building tools to retrieve information out
- Part 2 - Building the whole extraction system
Quick access to notebook: Dialogue_Data_Extraction_System.ipynb
Key concepts in the project
Function calling
- Single function calling
- Multiple function calling
- Parallel function calling
- Nested function calling
- No call
Function calling with external tools
- API interfacing
- Internal Python tool
Structured extraction
- Simple method
- Data class method
Function calling use cases
- Use case 1: extract structured data from unstructured data
- Use case 2: extract the most current data from web to self-learn and update
- Use case 3: retrieve insights from internal database
- Use case 4: generate valid JSON file
References
- Dataset on Hugging Face: SantiagoPG/customer_service_chatbot
- NexusRaven-V2 is used to proceed the function calling, which is from Nexusflow.
This project is supported by DeepLearning.AI and Nexusflow.
