metadata
license: apache-2.0
task_categories:
- text-generation
language:
- en
configs:
- config_name: default
data_files:
- split: train
path:
- shard_00000.parquet
- shard_00001.parquet
- shard_00002.parquet
- shard_00003.parquet
- shard_00004.parquet
Dataset
Mixed pretraining corpus - 1,000,000 rows, 5 shards.
Deduped (exact=False, fuzzy=False).