What problem does it solve?
This Skill addresses the challenge of manually packaging, validating, and submitting datasets to the AgentSociety platform, streamlining the process and ensuring dataset quality.
Core Features & Use Cases
- Dataset Creation and Initialization: Guides the user through creating dataset directories, initializing metadata, and setting up the structure for data files.
- Dataset Validation: Validates both directory and ZIP packages for correct format, metadata, and content, ensuring compatibility with the platform.
- Dataset Packaging: Packages directories into ZIP files for upload, with support for large files up to 5TB.
- Dataset Submission: Uploads the dataset to the AgentSociety platform and submits it for review.
- Use Case: If you have a collection of survey results, agent profiles, or simulation outputs and need to publish them on the AgentSociety platform, this Skill will automate the process.
Quick Start
Use the agentsociety-create-dataset skill to create a new dataset named 'my_dataset' and initialize it with the following command: agentsociety-create-dataset init my_dataset.