Sharegpt data fix#210

Open

surfiniaburger wants to merge 7 commits intometa-pytorch:mainfrom

surfiniaburger:sharegpt-data-fix

Commits on Nov 18, 2025

feat: Implement ShareGPT data generation and metadata recording

google-labs-jules[bot]
authored and
surfiniaburger
committed
fix: Align ShareGPT format with Hugging Face standards

google-labs-jules[bot]
authored and
surfiniaburger
committed
fix: Separate data and metadata for tool compatibility

google-labs-jules[bot]
authored and
surfiniaburger
committed
fix: Separate data and metadata and provide loading instructions

google-labs-jules[bot]
authored and
surfiniaburger
committed
fix: Restructure data for training and restore metadata logging

google-labs-jules[bot]
authored and
surfiniaburger
committed
refactor: Organize DIPG data generation scripts

google-labs-jules[bot]
authored and
surfiniaburger
committed
feat: Add robust ShareGPT data generation script

google-labs-jules[bot]
authored and
surfiniaburger
committed