nvidia/OpenCodeInstruct@8f3ba5bafe4d を複製し、カードを日本語で書く
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- .gitattributes +0 -1
- README.md +337 -0
- data/train-00000-of-00050.parquet +3 -0
- data/train-00001-of-00050.parquet +3 -0
- data/train-00002-of-00050.parquet +3 -0
- data/train-00003-of-00050.parquet +3 -0
- data/train-00004-of-00050.parquet +3 -0
- data/train-00005-of-00050.parquet +3 -0
- data/train-00006-of-00050.parquet +3 -0
- data/train-00007-of-00050.parquet +3 -0
- data/train-00008-of-00050.parquet +3 -0
- data/train-00009-of-00050.parquet +3 -0
- data/train-00010-of-00050.parquet +3 -0
- data/train-00011-of-00050.parquet +3 -0
- data/train-00012-of-00050.parquet +3 -0
- data/train-00013-of-00050.parquet +3 -0
- data/train-00014-of-00050.parquet +3 -0
- data/train-00015-of-00050.parquet +3 -0
- data/train-00016-of-00050.parquet +3 -0
- data/train-00017-of-00050.parquet +3 -0
- data/train-00018-of-00050.parquet +3 -0
- data/train-00019-of-00050.parquet +3 -0
- data/train-00020-of-00050.parquet +3 -0
- data/train-00021-of-00050.parquet +3 -0
- data/train-00022-of-00050.parquet +3 -0
- data/train-00023-of-00050.parquet +3 -0
- data/train-00024-of-00050.parquet +3 -0
- data/train-00025-of-00050.parquet +3 -0
- data/train-00026-of-00050.parquet +3 -0
- data/train-00027-of-00050.parquet +3 -0
- data/train-00028-of-00050.parquet +3 -0
- data/train-00029-of-00050.parquet +3 -0
- data/train-00030-of-00050.parquet +3 -0
- data/train-00031-of-00050.parquet +3 -0
- data/train-00032-of-00050.parquet +3 -0
- data/train-00033-of-00050.parquet +3 -0
- data/train-00034-of-00050.parquet +3 -0
- data/train-00035-of-00050.parquet +3 -0
- data/train-00036-of-00050.parquet +3 -0
- data/train-00037-of-00050.parquet +3 -0
- data/train-00038-of-00050.parquet +3 -0
- data/train-00039-of-00050.parquet +3 -0
- data/train-00040-of-00050.parquet +3 -0
- data/train-00041-of-00050.parquet +3 -0
- data/train-00042-of-00050.parquet +3 -0
- data/train-00043-of-00050.parquet +3 -0
- data/train-00044-of-00050.parquet +3 -0
- data/train-00045-of-00050.parquet +3 -0
- data/train-00046-of-00050.parquet +3 -0
- data/train-00047-of-00050.parquet +3 -0
.gitattributes
CHANGED
|
@@ -1,6 +1,5 @@
|
|
| 1 |
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
-
*.avro filter=lfs diff=lfs merge=lfs -text
|
| 4 |
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 5 |
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 6 |
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
|
|
|
| 1 |
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
*.arrow filter=lfs diff=lfs merge=lfs -text
|
|
|
|
| 3 |
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,337 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
pretty_name: nvidia/OpenCodeInstruct (mirror)
|
| 4 |
+
source_datasets:
|
| 5 |
+
- nvidia/OpenCodeInstruct
|
| 6 |
+
tags:
|
| 7 |
+
- laya
|
| 8 |
+
- laya-finetune
|
| 9 |
+
- mirror
|
| 10 |
+
- domain-code
|
| 11 |
+
- use-permissive
|
| 12 |
+
- genre-instruction
|
| 13 |
+
- label-score
|
| 14 |
+
- annot-llm
|
| 15 |
+
- annot-rule
|
| 16 |
+
- lang-en
|
| 17 |
+
language:
|
| 18 |
+
- en
|
| 19 |
+
task_categories:
|
| 20 |
+
- text-generation
|
| 21 |
+
size_categories:
|
| 22 |
+
- 1M<n<5M
|
| 23 |
+
configs:
|
| 24 |
+
- config_name: train
|
| 25 |
+
data_files:
|
| 26 |
+
- split: train
|
| 27 |
+
path: data/train-*
|
| 28 |
+
dataset_info:
|
| 29 |
+
- config_name: train
|
| 30 |
+
features:
|
| 31 |
+
- name: id
|
| 32 |
+
dtype: string
|
| 33 |
+
- name: input
|
| 34 |
+
dtype: string
|
| 35 |
+
- name: output
|
| 36 |
+
dtype: string
|
| 37 |
+
- name: domain
|
| 38 |
+
dtype: string
|
| 39 |
+
- name: generation_algorithm
|
| 40 |
+
dtype: string
|
| 41 |
+
- name: llm_judgement
|
| 42 |
+
dtype: string
|
| 43 |
+
- name: unit_tests
|
| 44 |
+
dtype: string
|
| 45 |
+
- name: tests_execution_status
|
| 46 |
+
dtype: string
|
| 47 |
+
- name: average_test_score
|
| 48 |
+
dtype: string
|
| 49 |
+
splits:
|
| 50 |
+
- name: train
|
| 51 |
+
num_bytes: 19066895906
|
| 52 |
+
num_examples: 5000000
|
| 53 |
+
download_size: 6861113102
|
| 54 |
+
dataset_size: 19066895906
|
| 55 |
+
---
|
| 56 |
+
|
| 57 |
+
# nvidia/OpenCodeInstruct(複製)
|
| 58 |
+
|
| 59 |
+
## データの内容
|
| 60 |
+
|
| 61 |
+
コード生成のための大規模な指示チューニング用の、コードの質問と回答の組のデータ。1 行が 1 組の質問と回答で、質問は `input`、LLM の応答は `output` に入る。self-instruct と evol-instruct の 2 つの方法で作られ、応答は LLM の評価と単体テストで採点されている(`llm_judgement`、`unit_tests`、`average_test_score`)。
|
| 62 |
+
|
| 63 |
+
## データの概要
|
| 64 |
+
|
| 65 |
+
| 項目 | 値 |
|
| 66 |
+
|---|---|
|
| 67 |
+
| 行数 | 約 4,969,570(1 構成・1 分割の合計) |
|
| 68 |
+
| 列数 | 9 |
|
| 69 |
+
| 容量 | 6.86 GB(ファイル 52 本) |
|
| 70 |
+
| ファイルの形式 | parquet 50 |
|
| 71 |
+
| 言語 | en |
|
| 72 |
+
| 公開 | public |
|
| 73 |
+
|
| 74 |
+
### 構成と分割
|
| 75 |
+
|
| 76 |
+
| 構成(config) | 分割(split)ごとの行数 | 列数 |
|
| 77 |
+
|---|---|---:|
|
| 78 |
+
| `train` | `train` 約 4,969,570 | 9 |
|
| 79 |
+
|
| 80 |
+
### 列(構成 `train`・分割 `train`)
|
| 81 |
+
|
| 82 |
+
| 列 | 型 | 先頭の行の値 |
|
| 83 |
+
|---|---|---|
|
| 84 |
+
| `id` | `string` | `4d61a897c0f86467f5c504a62a5f7282` |
|
| 85 |
+
| `input` | `string` | `You are given a list of 'n' tasks, each represen…` |
|
| 86 |
+
| `output` | `string` | `'''python def max_non_overlapping_tasks(tasks): …` |
|
| 87 |
+
| `domain` | `string` | `generic` |
|
| 88 |
+
| `generation_algorithm` | `string` | `self-instruct` |
|
| 89 |
+
| `llm_judgement` | `string` | `{"requirement_conformance": {"score": 5, "justif…` |
|
| 90 |
+
| `unit_tests` | `string` | `["\nassert max_non_overlapping_tasks([(1, 3), (2…` |
|
| 91 |
+
| `tests_execution_status` | `string` | `["pass", "pass", "fail", "pass", "pass", "pass",…` |
|
| 92 |
+
| `average_test_score` | `string` | `0.9` |
|
| 93 |
+
|
| 94 |
+
行数と列は HF のデータセットビューアの集計(元のデータセットの現在の版で数えたもの)。全データセットの一覧は [HYUKI5/dataset-catalog](https://huggingface.co/datasets/HYUKI5/dataset-catalog) にある。
|
| 95 |
+
|
| 96 |
+
## 複製について
|
| 97 |
+
|
| 98 |
+
[`nvidia/OpenCodeInstruct`](https://huggingface.co/datasets/nvidia/OpenCodeInstruct) をリビジョン `8f3ba5bafe4d6e8db46082cf7ae6741bc370604d` の時点でそのまま複製したもの。中身は変えていない。判断モデル Laya(`convaiinnovations/laya`)の微調整に使う候補として、調査メモ(`Laya 学習用データセット調査`)に載せたデータを集めている。
|
| 99 |
+
|
| 100 |
+
## 出典
|
| 101 |
+
|
| 102 |
+
- 元のデータセット: https://huggingface.co/datasets/nvidia/OpenCodeInstruct
|
| 103 |
+
- 複製したリビジョン: `8f3ba5bafe4d6e8db46082cf7ae6741bc370604d`(2025-04-28 19:08:02+00:00)
|
| 104 |
+
- 元のライセンス表記: `cc-by-4.0`
|
| 105 |
+
|
| 106 |
+
## 論文
|
| 107 |
+
|
| 108 |
+
- https://arxiv.org/abs/2504.04030
|
| 109 |
+
|
| 110 |
+
## アノテーションの方法
|
| 111 |
+
|
| 112 |
+
**データセットの作成日**
|
| 113 |
+
|
| 114 |
+
2025 年 1 月~2025 年 3 月
|
| 115 |
+
|
| 116 |
+
## 微調整に使うときの注意
|
| 117 |
+
|
| 118 |
+
- 調査メモの台帳での区分: A(許諾型のライセンスだけで、疑義の注記が無い)。区分は記載の文言で機械的に付けたもので、法的な判定ではない。学習に入れる前に元のカードとライセンスの原文で確かめる。
|
| 119 |
+
- Laya の入力は 2,048 token まで(ModernBERT-large / mmBERT)。長い本文は先頭から黙って切られるので、切れる行は事前に除くか、要約・切り出しをしてから入れる。
|
| 120 |
+
- 分割は記事・文書単位で行い、同じ本文から作った別の問いを学習側と較正側に分けない。温度と閾値は学習に使っていない分割で別々に当てる。
|
| 121 |
+
|
| 122 |
+
## 調査メモの記述
|
| 123 |
+
|
| 124 |
+
| メモ | 節 | ライセンス(記載) | 確認 | 注記 |
|
| 125 |
+
|---|---|---|---|---|
|
| 126 |
+
| `nvidia_nemotron.md` | 付録. 取得した nvidia 組織のデータセット全件の簡易一覧 (L368) | cc-by-4.0 | 確認済 | 確認印なし(各ファイルの凡例により確認済扱い) / 原文注記: × 数学・コード・科学・計測 |
|
| 127 |
+
|
| 128 |
+
## 元のカード(日本語訳)
|
| 129 |
+
|
| 130 |
+
以下は元のデータセットのカードの本文を日本語に訳したもの(機械翻訳)。訳に疑問があるときは、末尾の原文を確かめる。
|
| 131 |
+
|
| 132 |
+
---
|
| 133 |
+
|
| 134 |
+
# OpenCodeInstruct: コード LLM 向けの大規模な指示チューニングデータセット
|
| 135 |
+
|
| 136 |
+
|
| 137 |
+
## データセットの説明
|
| 138 |
+
|
| 139 |
+
OpenCodeInstruct を紹介する。500 万件の多様なサンプルからなる、最大規模のオープンアクセスの指示チューニングデータセットである。OpenCodeInstruct は教師あり微調整(SFT)のために設計されている。
|
| 140 |
+
|
| 141 |
+
- [技術報告書](https://arxiv.org/abs/2504.04030) - OpenCodeInstruct の手法と技術的な詳細。
|
| 142 |
+
- [GitHub リポジトリ](https://github.com/NVIDIA/NeMo-Skills) - SFT を行うために使われた完全なパイプライン。
|
| 143 |
+
|
| 144 |
+
このデータセットは商用・非商用のいずれにも利用できる。
|
| 145 |
+
|
| 146 |
+
|
| 147 |
+
## データセットの所有者
|
| 148 |
+
NVIDIA Corporation
|
| 149 |
+
|
| 150 |
+
|
| 151 |
+
## データセットの作成日
|
| 152 |
+
2025 年 1 月~2025 年 3 月
|
| 153 |
+
|
| 154 |
+
|
| 155 |
+
## ライセンス/利用条件
|
| 156 |
+
|
| 157 |
+
準拠する条件: このデータセットは Creative Commons Attribution 4.0 International License(CC BY 4.0)で提供される。ライセンス本文は https://creativecommons.org/licenses/by/4.0/legalcode で参照できる。
|
| 158 |
+
|
| 159 |
+
**データ開発者:** NVIDIA
|
| 160 |
+
|
| 161 |
+
### 利用ケース: <br>
|
| 162 |
+
コードの生成に特化させるために LLM を学習させる開発者。 <br>
|
| 163 |
+
|
| 164 |
+
### 公開日: <br>
|
| 165 |
+
04/28/2025 <br>
|
| 166 |
+
|
| 167 |
+
|
| 168 |
+
## 想定される用途
|
| 169 |
+
|
| 170 |
+
OpenCodeInstruct データセットは、オープンなモデルの改善を続けるためにコミュニティで利用されることを想定している。データはモデルの学習に自由に利用できる。**ただし、利用者は使用するデータセットごとに、そのライセンスが想定する目的に適するかを自ら確認する責任を負う**。
|
| 171 |
+
|
| 172 |
+
|
| 173 |
+
## データセットの特性
|
| 174 |
+
|
| 175 |
+
** データ収集の方法<br>
|
| 176 |
+
* [ハイブリッド: 自動、合成] <br>
|
| 177 |
+
|
| 178 |
+
** ラベル付けの方法<be>
|
| 179 |
+
* [ハイブリッド: 自動、合成] <br>
|
| 180 |
+
|
| 181 |
+
|
| 182 |
+
## データ形式
|
| 183 |
+
|Field|Type|Description|
|
| 184 |
+
|:---|:---|:---|
|
| 185 |
+
|id|string|質問ごとに一意の ID|
|
| 186 |
+
|input|string|入力となるコーディングの質問。|
|
| 187 |
+
|output|string|LLM の応答。|
|
| 188 |
+
|domain|string|"generic" または "algorithmic" のいずれか。|
|
| 189 |
+
|generation_algorithm|string|"self-instruct" または "evol-instruct" のいずれか。|
|
| 190 |
+
|llm_judgement|string|LLM による応答の評価を複数の基準で表した、JSON 辞書の文字列表現。|
|
| 191 |
+
|unit_tests|string|assert 文のリストの文字列表現。|
|
| 192 |
+
|tests_execution_status|string|"pass" または "fail" を示す文字列のリストの文字列表現。|
|
| 193 |
+
|average_test_score|float|合格したテストケースの割合。|
|
| 194 |
+
|
| 195 |
+
## 使い方
|
| 196 |
+
|
| 197 |
+
次の 2 行のコードでデータセットを読み込める。
|
| 198 |
+
|
| 199 |
+
```python
|
| 200 |
+
from datasets import load_dataset
|
| 201 |
+
opencodeinstruct = load_dataset("nvidia/OpenCodeInstruct", split="train")
|
| 202 |
+
```
|
| 203 |
+
|
| 204 |
+
|
| 205 |
+
## データセットの定量
|
| 206 |
+
|
| 207 |
+
- レコード数 - 500 万件のコードの質問と回答のペア。
|
| 208 |
+
- ダウンロードサイズ - 6.4 GB
|
| 209 |
+
|
| 210 |
+
|
| 211 |
+
## 倫理的な配慮:
|
| 212 |
+
|
| 213 |
+
NVIDIA は、信頼できる AI は共有の責任であると考えており、幅広い AI の応用を可能にするための方針と実践を定めている。利用規約に従ってダウンロード・利用する場合、開発者は、このモデルが関連する業界と用途の要件を満たし、想定外の製品の悪用に対応できるよう、社内のモデルチームと協力すべきである。
|
| 214 |
+
|
| 215 |
+
セキュリティ上の脆弱性や NVIDIA の AI に関する懸念は [こちら](https://www.nvidia.com/en-us/support/submit-security-vulnerability/) から報告すること。
|
| 216 |
+
|
| 217 |
+
|
| 218 |
+
## 引用
|
| 219 |
+
|
| 220 |
+
このデータが役に立った場合は、次を引用してほしい:
|
| 221 |
+
|
| 222 |
+
```
|
| 223 |
+
@article{ahmad2025opencodeinstruct,
|
| 224 |
+
title={OpenCodeInstruct: A Large-scale Instruction Tuning Dataset for Code LLMs},
|
| 225 |
+
author={Wasi Uddin Ahmad and Aleksander Ficek and Mehrzad Samadi and Jocelyn Huang and Vahid Noroozi and Somshubra Majumdar and Boris Ginsburg},
|
| 226 |
+
year={2025},
|
| 227 |
+
eprint={2504.04030},
|
| 228 |
+
archivePrefix={arXiv},
|
| 229 |
+
primaryClass={cs.CL},
|
| 230 |
+
url={https://arxiv.org/abs/2504.04030},
|
| 231 |
+
}
|
| 232 |
+
```
|
| 233 |
+
|
| 234 |
+
<details>
|
| 235 |
+
<summary>原文(英語)</summary>
|
| 236 |
+
|
| 237 |
+
# OpenCodeInstruct: A Large-scale Instruction Tuning Dataset for Code LLMs
|
| 238 |
+
|
| 239 |
+
|
| 240 |
+
## Dataset Description
|
| 241 |
+
|
| 242 |
+
We introduce OpenCodeInstruct, the largest open-access instruction tuning dataset, comprising 5 million diverse samples. OpenCodeInstruct is designed for supervised fine-tuning (SFT).
|
| 243 |
+
|
| 244 |
+
- [Technical Report](https://arxiv.org/abs/2504.04030) - Discover the methodology and technical details behind OpenCodeInstruct.
|
| 245 |
+
- [Github Repo](https://github.com/NVIDIA/NeMo-Skills) - Access the complete pipeline used to perform SFT.
|
| 246 |
+
|
| 247 |
+
This dataset is ready for commercial/non-commercial use.
|
| 248 |
+
|
| 249 |
+
|
| 250 |
+
## Dataset Owner(s)
|
| 251 |
+
NVIDIA Corporation
|
| 252 |
+
|
| 253 |
+
|
| 254 |
+
## Dataset Creation Date
|
| 255 |
+
January 2025 - March 2025
|
| 256 |
+
|
| 257 |
+
|
| 258 |
+
## License/Terms of Use
|
| 259 |
+
|
| 260 |
+
GOVERNING TERMS: This dataset is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0) available at https://creativecommons.org/licenses/by/4.0/legalcode.
|
| 261 |
+
|
| 262 |
+
**Data Developer:** NVIDIA
|
| 263 |
+
|
| 264 |
+
### Use Case: <br>
|
| 265 |
+
Developers training LLMs to specialize LLMs in code generation. <br>
|
| 266 |
+
|
| 267 |
+
### Release Date: <br>
|
| 268 |
+
04/28/2025 <br>
|
| 269 |
+
|
| 270 |
+
|
| 271 |
+
## Intended Usage
|
| 272 |
+
|
| 273 |
+
The OpenCodeInstruct Dataset is intended to be used by the community to continue to improve open models. The data may be freely used to train models. **However, for
|
| 274 |
+
each dataset a user elects to use, the user is responsible for checking if the dataset license is fit for the intended purpose**.
|
| 275 |
+
|
| 276 |
+
|
| 277 |
+
## Dataset Characterization
|
| 278 |
+
|
| 279 |
+
** Data Collection Method<br>
|
| 280 |
+
* [Hybrid: Automated, Synthetic] <br>
|
| 281 |
+
|
| 282 |
+
** Labeling Method<be>
|
| 283 |
+
* [Hybrid: Automated, Synthetic] <br>
|
| 284 |
+
|
| 285 |
+
|
| 286 |
+
## Dataset Format
|
| 287 |
+
|Field|Type|Description|
|
| 288 |
+
|:---|:---|:---|
|
| 289 |
+
|id|string|A unique id for each question|
|
| 290 |
+
|input|string|The input coding question.|
|
| 291 |
+
|output|string|LLM's response.|
|
| 292 |
+
|domain|string|Either "generic" or "algorithmic".|
|
| 293 |
+
|generation_algorithm|string|Either "self-instruct" or "evol-instruct".|
|
| 294 |
+
|llm_judgement|string|string representation of a JSON dictionary containing the LLM's evaluation of the response based on several criteria.|
|
| 295 |
+
|unit_tests|string|string representation of a list of assertion statements.|
|
| 296 |
+
|tests_execution_status|string|string representation of a list of strings indicating "pass" or "fail".|
|
| 297 |
+
|average_test_score|float|Fraction of test cases passed.|
|
| 298 |
+
|
| 299 |
+
## How to Use It
|
| 300 |
+
|
| 301 |
+
You can load the dataset with the following two lines of code.
|
| 302 |
+
|
| 303 |
+
```python
|
| 304 |
+
from datasets import load_dataset
|
| 305 |
+
opencodeinstruct = load_dataset("nvidia/OpenCodeInstruct", split="train")
|
| 306 |
+
```
|
| 307 |
+
|
| 308 |
+
|
| 309 |
+
## Dataset Quantification
|
| 310 |
+
|
| 311 |
+
- Record Count - 5 million coding question-answer pairs.
|
| 312 |
+
- Download Size - 6.4 GB
|
| 313 |
+
|
| 314 |
+
|
| 315 |
+
## Ethical Considerations:
|
| 316 |
+
|
| 317 |
+
NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.
|
| 318 |
+
|
| 319 |
+
Please report security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/).
|
| 320 |
+
|
| 321 |
+
|
| 322 |
+
## Citation
|
| 323 |
+
|
| 324 |
+
If you find the data useful, please cite:
|
| 325 |
+
```
|
| 326 |
+
@article{ahmad2025opencodeinstruct,
|
| 327 |
+
title={OpenCodeInstruct: A Large-scale Instruction Tuning Dataset for Code LLMs},
|
| 328 |
+
author={Wasi Uddin Ahmad and Aleksander Ficek and Mehrzad Samadi and Jocelyn Huang and Vahid Noroozi and Somshubra Majumdar and Boris Ginsburg},
|
| 329 |
+
year={2025},
|
| 330 |
+
eprint={2504.04030},
|
| 331 |
+
archivePrefix={arXiv},
|
| 332 |
+
primaryClass={cs.CL},
|
| 333 |
+
url={https://arxiv.org/abs/2504.04030},
|
| 334 |
+
}
|
| 335 |
+
```
|
| 336 |
+
|
| 337 |
+
</details>
|
data/train-00000-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:342757e0c6b706c8f68cf0a867ead5c417d0273726453dbcfc934f5c3c6ca891
|
| 3 |
+
size 133476478
|
data/train-00001-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2e822b1c8503683ab58aafe9eccbc546925f87bec45ced9b6040a43275e1d1d0
|
| 3 |
+
size 117050851
|
data/train-00002-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3fe342d7fb8616f4619a1f9183b71b42d9da3f5e68b9d1ca6d3afd6dbaee5037
|
| 3 |
+
size 115849397
|
data/train-00003-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:19aed4cc15ea04b3e9d842cb5bed9222fd4b0b84d59a206d4768642a27a82046
|
| 3 |
+
size 118088378
|
data/train-00004-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f25821f82620fa5687d914e20c5ced9639618597f8312b2839ab9fabc1ff6916
|
| 3 |
+
size 118220482
|
data/train-00005-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c60c3fd7160709a0ac6a3e23ff9d1fe8ff3d8810b1fea12d61741ca6710bd3bd
|
| 3 |
+
size 117103252
|
data/train-00006-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0d75dcebe8848f719c1feee7ec73752359a2e22a97596f86928462ea4cf75210
|
| 3 |
+
size 120968358
|
data/train-00007-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2839de0d6c4d1691710dbc63f43cd816e7c02a943f8e2c07087bbf5fa26172ce
|
| 3 |
+
size 113867035
|
data/train-00008-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:068b8d894e8a4c4ac98c50eb47677fa457a316481ca9e9fb35957c8a065cd0e3
|
| 3 |
+
size 121029590
|
data/train-00009-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c68c0fd487bdbfffef143ce898cc4af4f1dcae411099917d61484e7e47d331b7
|
| 3 |
+
size 127106336
|
data/train-00010-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c887eb84ebc90ff8b16db7f896d5b3386e4f315fea1ee7c97feece19da288ac7
|
| 3 |
+
size 132511783
|
data/train-00011-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:266733a7eca8c245defc555b21345de524c5a294c557f50919c4e2f0183167d1
|
| 3 |
+
size 154017621
|
data/train-00012-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:652d565412829057df94cfefe419a009b2b9fff468e7aeafcfadc31573d4985a
|
| 3 |
+
size 141178042
|
data/train-00013-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2dc8c89d6577b336fb0393436bd0b7706d2a7c1c1ab3a96358d36fa2c7afc83c
|
| 3 |
+
size 147068255
|
data/train-00014-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a2500ad51a54d0e62a96655a7dd468ff52f288ea3a02c7b380ded207d868a776
|
| 3 |
+
size 154560043
|
data/train-00015-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c610eec02f39206a87aa9a897f0b9b5764816d57b9c0a0cba3ab970127162c2
|
| 3 |
+
size 149728149
|
data/train-00016-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74b6ec1be10956f97fd19e9066d6740557f85176e023d8578e4b67a25b36d96a
|
| 3 |
+
size 139989167
|
data/train-00017-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7bf03f8606c18a93ca5e07b6af252e30de46b5609b2aa60133d8fe78c4c30154
|
| 3 |
+
size 154488894
|
data/train-00018-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4fb17d9a99408916ab6acd1d8ad63d7d1c918e4881c715bb6e16f1e1c95d6d04
|
| 3 |
+
size 131210694
|
data/train-00019-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a524fd4e65679b2f63c72cf34d34d0b9fde602cb91e47f30aa4cce20440d2120
|
| 3 |
+
size 138179442
|
data/train-00020-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c00b39b733ea7ae0141b7402b5e4c78645ceb72fe5479a5e669df20e0d74f992
|
| 3 |
+
size 153999706
|
data/train-00021-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:87834c5f7e71d161a2c8478add66ee72682791d5741a21bdf67031244f823d5f
|
| 3 |
+
size 149873743
|
data/train-00022-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b455bbb48193d0d05c59d1389c32e1d944ac454624f6716111968f1690c0886e
|
| 3 |
+
size 145842408
|
data/train-00023-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5ce7d1297f3e970b94209e2c7c997336cb39a6599fc14c7dda4d38f2fc14aa50
|
| 3 |
+
size 154577297
|
data/train-00024-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5a7d6cee09d56a30fdb6629d6e2ef4ad5f24d2e19c07c8ecb4d8e9e1df106c0a
|
| 3 |
+
size 150058021
|
data/train-00025-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e24337755630b41911f8303882c52aa46dfec3a499616110d38e5b34825d7a37
|
| 3 |
+
size 145021888
|
data/train-00026-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d4be4c94965f09ede41a91154240bf495a0508247bbc57b9186a072858de3e1b
|
| 3 |
+
size 149860137
|
data/train-00027-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f8987c9f00ff21c87bf50503c82152f2fd17684f7a42fd83632373ae395c691
|
| 3 |
+
size 145240836
|
data/train-00028-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a454c4dbaaada41a018398c3b831bf5395b4e29a20f6c966cce4134258eb310c
|
| 3 |
+
size 153829646
|
data/train-00029-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e50ae799468e3b41cb180155f70ce21ad6fcbdd6081c67cd1f97cc1aaf2bfa1a
|
| 3 |
+
size 146291069
|
data/train-00030-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1905b83021f20f14acd93a294a363c79e2e8ba67fd619401cafebfb30c03019d
|
| 3 |
+
size 144841412
|
data/train-00031-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af0e8dbea0889b57651e9f7ac15e74ceee39cb8c33cbcaacfd443443ba26993c
|
| 3 |
+
size 151902635
|
data/train-00032-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d6119262412a27bd035f2ae1d51d9fec1e8e761ebf08b7f2bc7739bb8f15770
|
| 3 |
+
size 148967375
|
data/train-00033-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:98a8b473a6dc7db7cb9c4d741c1789504b9935885876f30ada09729d562b6409
|
| 3 |
+
size 139833160
|
data/train-00034-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5c14f8ba1667e4ac1e84bd4f44ac89a34a5cbd74e64bfa34824925cba7d40aa
|
| 3 |
+
size 146366734
|
data/train-00035-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:14a94932c6a0a7560cbe1e4907e0185292f9aef290869c1fad4785ce38531178
|
| 3 |
+
size 141915562
|
data/train-00036-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64aa8064fc1edaa22f3af1e8e7673675077593d81aa5604672e0c370bdc41d1d
|
| 3 |
+
size 131648445
|
data/train-00037-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:730911f03f629bc86192408ae0f2be2a772c0d81b63a4295b77ae52038358f0b
|
| 3 |
+
size 135082977
|
data/train-00038-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:efbcaa6c2c8384931152302434fa23eca74fd43f048669c9ace6bc05ef11e076
|
| 3 |
+
size 127552271
|
data/train-00039-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8eabd81ca844e583ab9bfbff14abac78457b03d873cc595f19883bf0c1ddf22e
|
| 3 |
+
size 150642060
|
data/train-00040-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8b0d00ad0f20494c9d390face2e5dad7e34412479e501ef770ceb0a4e3c07e4
|
| 3 |
+
size 138595850
|
data/train-00041-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3fd29aca423b326062ba0408c5aeb6f9698ed63876a707dc4ecc85ce789b0a6b
|
| 3 |
+
size 127448792
|
data/train-00042-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddd46d5c10495eb1ef526e9a41cb8ac0aead98962ee49cde82a3d7d32fa8c91d
|
| 3 |
+
size 131590828
|
data/train-00043-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5070283c602d44e73a1fc79df6030d6476f3e403cba2b1159d32f6d89943cd79
|
| 3 |
+
size 127362865
|
data/train-00044-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2dc859fce71fe1387dee5c6d3c0b5ed32ac33fde4434dafa0cb59b859ade26ae
|
| 3 |
+
size 127388210
|
data/train-00045-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cd4f543fcfeb59bfa9e8c8e36c56923741c995ad14f0995c4f70e9e8c3fc2e01
|
| 3 |
+
size 127527968
|
data/train-00046-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16d45b54b84d11bd4061ab7f370927924813f8636077c6c6fe51f65c5081745b
|
| 3 |
+
size 131222516
|
data/train-00047-of-00050.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad78ddbba73c0baa3d741bb4873d88bcb5436469f000220b035993407fed1de2
|
| 3 |
+
size 127413481
|