OpenCodeInstruct
← All datasets
Author: hf-mirror · License: cc-by-4.0 · Format: parquet
0 downloads · 0 stars · 0 mirrors · 6,861,113,102 · 50 files
NVIDIA's 5.7 GB coding-SFT corpus: synthetic programming problems with solutions, tests, and execution feedback.
Tags: coding · sft · nvidia · synthetic
Description
NVIDIA's open instruction-tuning corpus for code LLMs (OpenCodeInstruct): 50 parquet shards (~5.7 GB) of synthetic coding problems with generated solutions, test cases, and execution feedback, built with Llama-based teachers for SFT. CC-BY-4.0.
Download
Direct download: https://huggingface.co/datasets/nvidia/OpenCodeInstruct/tree/main
Or via the API: curl -L -o file "https://huggingface.co/datasets/nvidia/OpenCodeInstruct/tree/main"
Files
- data/train-00000-of-00050.parquet (133,476,478)
- data/train-00001-of-00050.parquet (117,050,851)
- data/train-00002-of-00050.parquet (115,849,397)
- data/train-00003-of-00050.parquet (118,088,378)
- data/train-00004-of-00050.parquet (118,220,482)
- data/train-00005-of-00050.parquet (117,103,252)
- data/train-00006-of-00050.parquet (120,968,358)
- data/train-00007-of-00050.parquet (113,867,035)
- data/train-00008-of-00050.parquet (121,029,590)
- data/train-00009-of-00050.parquet (127,106,336)
- data/train-00010-of-00050.parquet (132,511,783)
- data/train-00011-of-00050.parquet (154,017,621)
- data/train-00012-of-00050.parquet (141,178,042)
- data/train-00013-of-00050.parquet (147,068,255)
- data/train-00014-of-00050.parquet (154,560,043)
- data/train-00015-of-00050.parquet (149,728,149)
- data/train-00016-of-00050.parquet (139,989,167)
- data/train-00017-of-00050.parquet (154,488,894)
- data/train-00018-of-00050.parquet (131,210,694)
- data/train-00019-of-00050.parquet (138,179,442)
- data/train-00020-of-00050.parquet (153,999,706)
- data/train-00021-of-00050.parquet (149,873,743)
- data/train-00022-of-00050.parquet (145,842,408)
- data/train-00023-of-00050.parquet (154,577,297)
- data/train-00024-of-00050.parquet (150,058,021)
- data/train-00025-of-00050.parquet (145,021,888)
- data/train-00026-of-00050.parquet (149,860,137)
- data/train-00027-of-00050.parquet (145,240,836)
- data/train-00028-of-00050.parquet (153,829,646)
- data/train-00029-of-00050.parquet (146,291,069)
- data/train-00030-of-00050.parquet (144,841,412)
- data/train-00031-of-00050.parquet (151,902,635)
- data/train-00032-of-00050.parquet (148,967,375)
- data/train-00033-of-00050.parquet (139,833,160)
- data/train-00034-of-00050.parquet (146,366,734)
- data/train-00035-of-00050.parquet (141,915,562)
- data/train-00036-of-00050.parquet (131,648,445)
- data/train-00037-of-00050.parquet (135,082,977)
- data/train-00038-of-00050.parquet (127,552,271)
- data/train-00039-of-00050.parquet (150,642,060)
- data/train-00040-of-00050.parquet (138,595,850)
- data/train-00041-of-00050.parquet (127,448,792)
- data/train-00042-of-00050.parquet (131,590,828)
- data/train-00043-of-00050.parquet (127,362,865)
- data/train-00044-of-00050.parquet (127,388,210)
- data/train-00045-of-00050.parquet (127,527,968)
- data/train-00046-of-00050.parquet (131,222,516)
- data/train-00047-of-00050.parquet (127,413,481)
- data/train-00048-of-00050.parquet (124,941,534)
- data/train-00049-of-00050.parquet (138,581,429)
Added 2026-09-27. Platform: Distillio — the open source training data collective.