---
configs:
- config_name: Markdown
  default: true
  data_files:
  - split: train
    path: data/Markdown/train/*.parquet
- config_name: AsciiDoc
  data_files:
  - split: train
    path: data/AsciiDoc/train/*.parquet
- config_name: Org
  data_files:
  - split: train
    path: data/Org/train/*.parquet
- config_name: Pod
  data_files:
  - split: train
    path: data/Pod/train/*.parquet
- config_name: RMarkdown
  data_files:
  - split: train
    path: data/RMarkdown/train/*.parquet
- config_name: Roff
  data_files:
  - split: train
    path: data/Roff/train/*.parquet
- config_name: Roff Manpage
  data_files:
  - split: train
    path: data/Roff Manpage/train/*.parquet
- config_name: TeX
  data_files:
  - split: train
    path: data/TeX/train/*.parquet
- config_name: Texinfo
  data_files:
  - split: train
    path: data/Texinfo/train/*.parquet
- config_name: Text
  data_files:
  - split: train
    path: data/Text/train/*.parquet
- config_name: Textile
  data_files:
  - split: train
    path: data/Textile/train/*.parquet
- config_name: Typst
  data_files:
  - split: train
    path: data/Typst/train/*.parquet
- config_name: Vim Help File
  data_files:
  - split: train
    path: data/Vim Help File/train/*.parquet
- config_name: Wikitext
  data_files:
  - split: train
    path: data/Wikitext/train/*.parquet
- config_name: reStructuredText
  data_files:
  - split: train
    path: data/reStructuredText/train/*.parquet
license: odc-by
task_categories:
- text-generation
---

# stack-v3-docs


| Language | Samples |
| --- | ---: |
| Markdown | 199,003,312 |
| TeX | 5,139,532 |
| reStructuredText | 3,067,032 |
| Text | 2,752,971 |
| RMarkdown | 1,959,373 |
| Org | 1,043,577 |
| AsciiDoc | 840,637 |
| Roff | 570,452 |
| Wikitext | 455,210 |
| Roff Manpage | 406,048 |
| Textile | 64,706 |
| Typst | 43,944 |
| Pod | 40,833 |
| Vim Help File | 38,634 |
| Texinfo | 21,402 |

- Exact duplicate samples were removed after extraction.

## Licensing information

These subsets are derived from [The Stack v3 train dataset](https://huggingface.co/datasets/HuggingFaceCode/stack-v3-train), which is released under the [ODC-By 1.0 license](https://opendatacommons.org/licenses/by/1-0/). Individual source files may have different original licenses; their use must comply with those terms, including attribution where applicable. Repository provenance and per-file license metadata (`repo_path`, `commit_id`, `license_type`, and `detected_licenses`) are preserved in each sample. License detection may be incomplete or inaccurate.