Dataset · starcoderdata

StarCoderData

Builder: BigCode
Availability: gated · checked 2026-09-24 · source
HF API reports gated: auto (approval step); the card text could not be read without accepting terms.

Raw record: /data/datasets/starcoderdata.json

Fields

id

starcoderdata

identifiers

huggingface

bigcode/starcoderdata

builder

BigCode

release_date

2023-03-30partial · source

note:  HF dataset repo creation date, not the builder's own release announcement.

availability

gated · checked 2026-09-24 · source

note:  HF API reports gated: auto (approval step); the card text could not be read without accepting terms.

content

TinyLlama paper: 'the training data of StarCoder ... contains code data in 86 programming languages. In addition to code, it also includes GitHub issues and text-code pairs that involve natural languages.' StarCoder paper: StarCoderBase trained on 1 trillion tokens from The Stack v1.2 (permissively licensed code, 44 opt-outs at processing time).partial · source

note:  Primary card was gated; description rests on the TinyLlama and StarCoder papers.

primary_sources

https://arxiv.org/abs/2305.06161

https://huggingface.co/datasets/bigcode/starcoderdata

record_history

date:2026-09-24 · change:created from primary sources (tranche 2 dataset records) · by:wilson-pruitt + claude ·

Parents

No edges recorded.

Children

Training data

• ← trained_on StarCoderBase declared source Dataset card (public API description): 'This is the dataset used for training StarCoder and StarCoderBase. It contains 783GB of code in 86 programming languages ... approximately 250 Billion tokens.' NOTE the 250B-token figure differs from the 1T tokens seen in training (multiple epochs per the paper); the dataset file itself is gated.
• ← trained_on TinyLlama-1.1B (intermediate-step-1431k-3T) declared source Card metadata `datasets: bigcode/starcoderdata`; paper: 'the training data of StarCoder', code-related samples only. `starcoderdata` is gated.
• ← trained_on StarCoder declared source Dataset card (public API description): 'the dataset used for training StarCoder and StarCoderBase'. StarCoder's own fine-tuning used the Python subset of that training data (paper sec. 5.6).

Read in

No station on the reading path has touched this record yet.