Not Just Talk: Verbi, the Voice Assistant That Acts
Updated: November 19, 2024
Summary
The video introduces WBY, an open-source assistant capable of automating tasks with the use of llms and community contributions. It demonstrates the workflow of WBY, from input retrieval to generating responses using external tools or APIs. The video also shows how to set up and customize WBY's functionalities locally, including function calling and supported actions like calendar events and weather information. Viewers can access the code on GitHub to start enhancing their own virtual assistant experience.
Calendar and Tasks Overview
The user checks the calendar and tasks, finding a team meeting, dentist appointment, dinner plans, pending tasks including project proposal and buying groceries.
Introduction to WBY Assistant
Introduction to WBY, an open-source wise assistant capable of taking actions on behalf of the user with the help of Community and function calling abilities of modern llms.
Workflow of WBY
Explanation of WBY's workflow, including getting input from the user, deploying a speech to text model, using an llm for decision-making, leveraging external tools or APIs, generating responses, and playing back the generated speech.
Local Models Implementation
Demonstration of how to replace steps with local models, availability of code on GitHub, cloning the agent branch, setting up a virtual environment, installing required packages, and setting up API keys for different providers locally.
Function Calling and Tool Selection
Explanation of how function calling works in WBY, including the llm selecting available tools, determining the need for a specific tool, making external API calls, generating responses, and updating the main code for incorporating function calling.
Actions Supported and Tool Implementation
Overview of supported actions in WBY, implementation of tools for calendar events, email access, task listing, and weather information using dummy data and specific versions of models.
System Prompt and Python Execution
Details on the system prompt in the main file, generating thoughtful responses, setting current date, running the assistant using Python, and testing with user queries for party planning and weather forecast.
FAQ
Q: What is the purpose of WBY as described in the file?
A: WBY is an open-source wise assistant designed to take actions on behalf of the user by leveraging the Community's help and the function calling abilities of modern language models.
Q: Can you explain the workflow of WBY as outlined in the file?
A: The workflow of WBY involves getting input from the user, deploying a speech to text model, using language models for decision-making, utilizing external tools or APIs, generating responses, and playing back the generated speech.
Q: How is the process of replacing steps with local models described in the file?
A: The file describes the process of replacing steps with local models by making the code available on GitHub, cloning the agent branch, setting up a virtual environment, installing required packages, and setting up API keys for different providers locally.
Q: What is function calling in WBY, and how does it work according to the document?
A: Function calling in WBY involves the language model selecting available tools, determining the need for a specific tool, making external API calls, generating responses, and updating the main code to incorporate function calling.
Q: What are some of the supported actions in WBY as mentioned in the file?
A: WBY supports actions related to calendar events, email access, task listing, and weather information using dummy data and specific versions of models.
Q: How is the system prompt in the main file described?
A: The system prompt in the main file is detailed as generating thoughtful responses, setting the current date, running the assistant using Python, and testing with user queries for party planning and weather forecast.
Get your own AI Agent Today
Thousands of businesses worldwide are using Chaindesk Generative
AI platform.
Don't get left behind - start building your
own custom AI chatbot now!