AI Briefing#dataset
The latest AI and developer news about #dataset, with the original source and a short summary.
The latest AI and developer news about #dataset, with the original source and a short summary.
An Indic corpus spanning 11 languages and 8.4B tokens has been released on Hugging Face.
AllenAI has released the MolmoAct2 robot control model along with fine-tuned versions.
legalize-kr, a project that archives South Korean statutes and case law as Markdown and Git history, has been updated with support for an MCP server and Agent Skill.
A pseudo-annotation pipeline was proposed that automatically generates annotation candidates based on sign language video and English.
DharmaOCR released specialized 3B and 7B SLMs along with cost and performance benchmarks.
A multimodal prompt injection dataset v5 with 503,358 samples has been released.
GM-SEUS v2 has released 3.42 million panels and a new rooftop array.
MIT and IMO have released MathNet, spanning 40 countries and 4 decades.
About 1,200 of the roughly 5,300 papers accepted to ICLR 2026 provide public code, data, or demos.
They trained a fast and accurate 6-language OCR model on 12.2 million synthetic data images.
A dataset of 100,000 CoT samples for fine-tuning local reasoning models has been released.
A large-scale dataset containing over 20 million Indian legal documents, citation graphs, and vector embeddings has been released.
It archives South Korean statutes and precedents as Markdown and Git history, providing a dataset optimized for LLM training and RAG.
A new benchmark called Alyah has been released to measure Arabic LLMs' understanding of the Emirati dialect.
HuggingFace improved the streaming efficiency of the `datasets` library, speeding up large-scale data training.
NVIDIA has released a synthetic persona dataset of 21 million people that reflects India's cultural and linguistic characteristics.
NVIDIA has released a large-scale synthetic persona dataset reflecting Japan's demographics and culture.
Hugging Face has released LeRobotDataset v3.0 for handling large-scale robot learning data.
SandboxAQ has released SAIR, a dataset of 5.24 million 3D protein-ligand structures for predicting drug efficacy.
Hugging Face has launched AI Sheets, a no-code tool that lets you build and transform datasets using open models.
Hugging Face is building a community-driven LeRobot dataset ecosystem to secure generalization for robots.
The Open R1 team has released OlympicCoder, a coding model that outperforms Claude 3.7 Sonnet, along with new datasets.
Hugging Face and Entalpic have unveiled LeMaterial, an open source project to accelerate materials research, along with the unified dataset LeMat-Bulk.
The Hugging Face community has released an open-source preference dataset for training text-to-image generation models.