README.md

metadata

pretty_name: MMVistaReason-SFT-528K
configs:
  - config_name: 49K
    data_files:
      - split: train
        path: MMVistaReason-SFT-49K.parquet
  - config_name: 124K
    data_files:
      - split: train
        path: MMVistaReason-SFT-124K.parquet
  - config_name: 355K
    data_files:
      - split: train
        path: MMVistaReason-SFT-355K.parquet

MMVistaReason-SFT-528K

MMVistaReason supervised fine-tuning data, containing 49K, 124K, and 355K teacher-distilled subsets in Parquet format.