SKU SFT-2305 · Sold by External
OpenCodeReasoning (OCR-1)
Product specifications
| SKU | SFT-2305 |
|---|---|
| Data type | SFT dataset |
| Volume | 735255 rows |
| Size on disk | Not published by source |
| Format | Parquet (Hugging Face) |
| Access model | PUBLIC LICENSE |
| Pricing | Free · open-source license |
| Quality score | — |
| License | CC-BY-4.0 |
NVIDIA's OpenCodeReasoning is a synthetic reasoning SFT dataset of 735,255 Python samples spanning 28,319 unique competitive-programming questions (TACO, APPS, CodeContests, CodeForces, CodeChef, AtCoder), with reasoning traces generated by DeepSeek-R1. Delivered in two configs: split_0 (567,850 rows) and split_1 (167,405 rows). It is code reasoning (chain-of-thought), not API/function-calling data.