Canonical large artifacts for reproducing real-data Acta v0.2 benchmarks.
The converters, schema mappings, commands, checksums, validation procedures,
and measured results live in the
Acta GitHub repository.
Contents
bts_flight/v0.2/
bts_flights_15750000.parquet: 15,750,000 rows, 51 columns, 381,644,436 bytes;
bts_flights_15750000.acta: adaptive Acta v0.2 output using Zstandard level 6, 325,340,952 bytes;
manifest.json: source… See the full description on the dataset page: https://huggingface.co/datasets/lu-chengass/acta-data.
