language:
- en
A cleaned version of the stack.
Based on the language distribution in the original data, only the following high-resource programming languages are retained: java, javascript, php, python, c-sharp, typescript, c, cpp, go, html, ruby, kotlin, shell, rust. Non-conventional programming languages such as markdown and json are removed.
All non-English files are removed.
Low-quality files are heuristically removed.
This version has also been decontaminated with respect to the following benchmarks based on n-gram overlap:
- GLUE (dev set of SST-2, CoLA, QQP, WNLI, RTE, QNLI, MNLI; test set of MPRC)
- SIQA, PIQA, QASC, CSQA, HellaSWAG (all dev set)
- CONLL 2003
- BLIMP
- MAIN
- BoolQ (dev set)
- WinoGrande (dev set)
- ANLI (test set)
- ARC easy and challenge (test set)
- RACE middle and high (test set)
- MMLU (dev, val, and test sets)
- MATH, GSM8K (test set)
- HumanEval (test set)
- MBPP (all 974 questions)
- GPQA (diamond)
Dataset Statistics
Number of samples: 94,940,432.
Size of downloaded parquet files: 142G.
Other Languages
In the zh branch, samples containing Chinese text (e.g. comments, strings) are also collected (4,124,610 samples, 6.1G parquet files). They are processed by the same pipeline and do not overlap with samples in the main branch. You can download them by specifying the --revision argument:
huggingface-cli download --repo-type dataset Geralt-Targaryen/stack --revision zh --local-dir .