coder-pretrain-60gb / README.md
Aniket200325's picture
Upload README.md with huggingface_hub
4e28997 verified
|
Raw
History Blame Contribute Delete
947 Bytes
metadata
license: other
language:
  - en
size_categories:
  - 10B<n<100B
task_categories:
  - text-generation
tags:
  - code
  - pretraining
configs:
  - config_name: default
    data_files:
      - split: train
        path: final/train/*.parquet
      - split: validation
        path: final/validation/*.parquet

Coding LLM Pretraining Corpus

Built with build_pretrain_dataset.ipynb + finalize_local.py (DuckDB).

Composition

Source Collected rows Collected size Final rows Final size
code 8,415,197 30.00 GB 8,413,850 29.88 GB
web 3,492,492 15.00 GB 3,390,372 14.50 GB
math 659,720 3.00 GB 658,089 2.95 GB
wiki 519,584 2.00 GB 519,535 1.99 GB
docs 4,504,069 10.00 GB 4,289,545 9.74 GB

Total final corpus: 59.06 GB of raw text (17,271,391 documents) across 159 train shards and 1 validation shards. 91,189 documents were removed by dedup/decontamination.