Skip to content

Inconsistency between github's data.jsonl and huggingface ToolScale data #18

@shirley-wu

Description

@shirley-wu

Hello, I notice that you released your tool calling data both at https://github.com/NVlabs/ToolOrchestra/blob/main/data/data.jsonl and https://huggingface.co/datasets/nvidia/ToolScale, but the two sets of data seem largely different. Is data.jsonl intended to also be the ToolScale dataset? And could you clarify which data is the correct one to reproduce your work? Thanks!

Also, may I also ask what's your data mixture between ToolScale and GeneralThoughts-430k-filtered?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions