Dataset Curation

17 summarised stories about Dataset Curation, each linking back to the original source. Browse all topics →

Tuesday, 11 March 2025

Open R1: Update #3

Hugging Face Blog 1 year ago

Researchers released CodeForces-CoTs, a dataset of nearly 100,000 chain-of-thought samples distilled from DeepSeek-R1 for competitive programming in C++ and Python, along with OlympicCoder models fine-tuned on this data. OlympicCoder-32B outperformed Claude 3.7 Sonnet and achieved higher scores than o1-mini and DeepSeek-R1 on the IOI 2024 benchmark, despite being 7 billion to 32 billion parameters. The release enables open-source development of code reasoning models and provides a new fully-verifiable benchmark from International Olympiad in Informatics problems, addressing limitations in existing competitive programming datasets that only include partial test cases.

Benchmarking Code Generation Dataset Curation Open Source

LeRobot goes to driving school: World’s largest open-source self-driving dataset

Hugging Face Blog 1 year ago

Hugging Face and Yaak released L2D, an open-source self-driving dataset containing 90+ terabytes of multimodal driving data collected from 60 vehicles in 30 German cities over three years. The dataset comprises 1 million episodes spanning 5,000+ hours of driving, captured with six RGB cameras, GPS, IMU, and vehicle control signals (steering, gas, brake). The data includes both expert driving instructor policies and suboptimal student driver policies with natural language instructions, enabling training of end-to-end autonomous driving models through imitation learning.

Autonomous Vehicles Dataset Curation Hugging Face Open Source