hestia-chickens2.mp4
The utbots_llama package integrates Retrieval-Augmented Generation (RAG) with Meta's LLaMA language model within the Robot Operating System (ROS) framework. This enables robots to answer context-based questions by retrieving relevant information and generating coherent responses. Features
Contextual Question Answering: Enhances robot interactions by providing accurate, context-aware responses.
Seamless ROS Integration: Ensures smooth communication between the RAG system and other ROS components.
Modular Design: Facilitates easy customization and extension of functionalities.
Installation
Clone the Repository:
git clone https://github.com/UtBotsAtHome-UTFPR/utbots_llama.git
Install Dependencies:
cd utbots_llama pip install -r requirements.txt Needs to perform: https://gist.github.com/defulmere/8b9695e415a44271061cc8e272f3c300 to work.
Build the Package:
catkin_make
Usage
Launch the RAG Node:
roslaunch task_manager llama_qa.launch
Interacting with the Robot: Use ROS topics or services to send questions and receive responses.
Configuration
Model Settings: Adjust parameters in the config/model.yaml file to fine-tune the LLaMA model.
Data Sources: Specify knowledge base locations in the config/data_sources.yaml file.
Contributing
We welcome contributions! Please fork the repository and submit a pull request with your changes. License
This project is licensed under the MIT License.