🔧 Retrieving useful tools from a large-scale toolset is an important step for Large language model (LLMs) in tool learning. This project (ToolRet) contribute to (i) the first comprehensive tool retrieval benchmark to systematically evaluate existing information retrieval (IR) models on tool retrieval tasks; and (ii) a large-scale training dataset to optimize the expertise of IR models on this tool retrieval task.
This ToolRet-Tools contains the toolset corpus of our tool retrieval benchmark.… See the full description on the dataset page:
https://huggingface.co/datasets/mangopy/ToolRet-Tools.
