Datasets:

Beetle-Data
/

ca-100M-pretok

Formats:

parquet

Size:

Libraries:

Dataset card Data Studio Files Files and versions

xet

Community

ca-100M-pretok / README.md

suchirsalhan

Add data/ca_clean_00000-00.parquet

31077db verified about 2 months ago

preview code

Raw

History Blame Contribute Delete

292 Bytes

metadata

configs:
  - config_name: default
    data_files: data/*.parquet

Beetle-Data/ca-100M-pretok

Pretokenized chunks (input_ids = 513-token packed sequences, no cross-document bleeding). Sharded incrementally; the marker data/_finalized.json is committed once all parts are uploaded.