Divide-Then-Aggregate

📃 Paper • 🌐 Project Page • 🤗 Dataset (DTA-Tool) • 🤗 Model (DTA-Llama)

This project is from the paper "Divide-Then-Aggregate: An Efficient Tool Learning Method via Parallel Tool Invocation". It makes a large language model capable of invoking tools and can parallel invoke multiple tools within a single round.

How to use

In this project, there are two sub-projects naming tool_llama_client and tool_llama_server. The tool_llama_client contains training scrips and inference parts, while we keep the model prediction part in tool_llama_server so that we can deploy the model in a machine with powerful GPUs and run the inference client script in a machine with network connection to openAI.

Please following the Readme.md in each sub-project.

• tool_llama_client: training, running the inference client and eval the inference results.
• tool_llama_server: deploying the trained model, such as llama2-7b

License

This repository's code is under Apache License 2.0. Many codes are based on ToolBench and StableToolBench with Apache-2.0 License.