datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
P3
Dataset Card for P3
Dataset Summary
P3 (Public Pool of Prompts) is a collection of prompted English datasets covering a diverse set of NLP tasks. A prompt is the combination of an input template and a target template. The templates are functions mapping a data example into natural language for the input and target sequences. For example, in the case of an NLI dataset, the data example would include fields for Premise, Hypothesis, Label. An input template would be If… See the full description on the dataset page: https://huggingface.co/datasets/bigscience/P3.stack-v3-train
🥞 The Stack v3
What is it?
What is being released
How to download and use it
Dataset statistics
Dataset structure
Dataset creation
Considerations for using the data
Additional information
What is it?
The Stack v3 is the largest, most up-to-date open dataset of source code, crawled directly from GitHub and built to pre-train code LLMs with full-repository context. It is the successor to The Stack v2 and, like its predecessor, is released to make the training… See the full description on the dataset page: https://huggingface.co/datasets/HuggingFaceCode/stack-v3-train.agibot_alpha_v30This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "AgiBot_A2D",
"total_episodes": 28122,
"total_frames": 47613574,
"total_tasks": 30,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:28122"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/cadene/agibot_alpha_v30.tulu-3-sft-mixture
Tulu 3 SFT Mixture
Note that this collection is licensed under ODC-BY-1.0 license; different licenses apply to subsets of the data. Some portions of the dataset are non-commercial. We present the mixture as a research artifact.
The Tulu 3 SFT mixture was used to train the Tulu 3 series of models.
It contains 939,344 samples from the following sets:
CoCoNot (ODC-BY-1.0), 10,983 prompts (Brahman et al., 2024)
FLAN v2 via ai2-adapt-dev/flan_v2_converted, 89,982 prompts (Longpre et… See the full description on the dataset page: https://huggingface.co/datasets/allenai/tulu-3-sft-mixture.multispan_quoref3DCode
Project page
Paper
Code
3dcodebench.com
arXiv:2606.01057
gaoypeng/3dcodebench
News
[06/01/2026] Paper released on arXiv: 3DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Code.
Note. This is an open-source reproduction of 3DCodeBench.
⚠️ Under final check. The 3DCodeData/ code is still undergoing final
quality review and may contain occasional issues (non-executable scripts, mismatched
captions/renders, or imperfect geometry). If you run… See the full description on the dataset page: https://huggingface.co/datasets/YipengGao/3DCode.Ultra-FineWeb-L3
Ultra-FineWeb-L3
📜 Ultra-FineWeb Technical Report |
📦 UltraData Collection |
🌐 UltraData |
🤗 MiniCPM5 Series
English |
中文
📚 Introduction
Ultra-FineWeb-L3 is the L3 refined data for general high-quality web data within UltraData's L0-L4 tiered data management framework. Moving beyond L2 quality selection, it transforms high-value web corpora into structured, high-learnability training data with clearer reasoning signals and richer educational… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/Ultra-FineWeb-L3.BVD-I-300M-URLs
LAION-BVD - 300M Video Frame URLs
This repository contains the URLs for ~300 million keyframes extracted from publicly available web videos. No image data is included, only the source video URL and the frame timestamp needed to reproduce each frame.
Frames were extracted from BVD-RAW and cover YouTube, Dailymotion, and Vimeo content.
Dataset structure
Column
Type
Description
webpage_url
string
URL of the source video
frame_pts_time
float
Presentation… See the full description on the dataset page: https://huggingface.co/datasets/laion/BVD-I-300M-URLs.TxT360-v2
TxT360-v2
Dataset Description
Pre-training sources for the K2 Horizon training data release. This repository is part of the K2 Horizon collection.
The repository is organized into multiple subsets. Every subset has a train split backed by Parquet shards.
K2 Horizon Dataset Series
Dataset repository
Focus
Subsets
IFM/TxT360-v2
Web and question-answering text
3
IFM/Code-Reasoning
Code reasoning and task synthesis
7
IFM/Math-Reasoning… See the full description on the dataset page: https://huggingface.co/datasets/IFM/TxT360-v2.OpenThoughts3-1.2M
paper |
dataset |
model
[!NOTE]
We have released a paper for OpenThoughts! See our paper here.
OpenThoughts3-1.2M
Open-source state-of-the-art reasoning dataset with 1.2M rows. 🚀
OpenThoughts3-1.2M is the third iteration in our line of OpenThoughts datasets, building on our previous OpenThoughts-114k and OpenThoughts2-1M.
This time around, we scale even further and generate our dataset in a much more systematic way -- OpenThoughts3-1.2M is the result of a… See the full description on the dataset page: https://huggingface.co/datasets/open-thoughts/OpenThoughts3-1.2M.symile-m3
Dataset Card for Symile-M3
Symile-M3 is a multilingual dataset of (audio, image, text) samples. The dataset is specifically designed to test a model's ability to capture higher-order information between three distinct high-dimensional data types: by incorporating multiple languages, we construct a task where text and audio are both needed to predict the image, and where, importantly, neither text nor audio alone would suffice.
Paper: https://arxiv.org/abs/2411.01053
GitHub:… See the full description on the dataset page: https://huggingface.co/datasets/arsaporta/symile-m3.tulu-3-sft-personas-instruction-following
Dataset Descriptions
This dataset contains 29980 examples and is synthetically created to enhance model's capabilities to follow instructions precisely and to satisfy user constraints. The constraints are borrowed from the taxonomy in IFEval dataset.
To generate diverse instructions, we expand the methodology in Ge et al., 2024 by using personas. More details and exact prompts used to construct the dataset can be found in our paper.
Curated by: Allen Institute for AI
Paper: TBD… See the full description on the dataset page: https://huggingface.co/datasets/allenai/tulu-3-sft-personas-instruction-following.sun397
SUN397 dataset
The database contains 397 categories subset from the SUN dataset for Scene Recognition used in the following paper.
The number of images varies across categories, but there are at least 100 images per category, and 108,754 images in total.
All images are in jpg format. The images provided here are for research purposes only.
The file ClassName.txt contains the name list for the 397 categories.
Please cite the following paper if you use this dataset in your research.… See the full description on the dataset page: https://huggingface.co/datasets/tanganke/sun397.genshin-voice
Genshin Voice
Genshin Voice is a dataset of voice lines from the popular game Genshin Impact.
Hugging Face 🤗 Genshin-Voice
ModelScope Genshin-Voice
Per-speaker downloads are grouped by language and ZIP size. Browse every archive in the ZIP index.
Last update at 2026-08-13
654252 wavs
7291 without speaker (1%)
52693 without transcription (8%)
1088 without inGameFilename (0%)
Dataset Details
Dataset Description
The dataset contains voice lines… See the full description on the dataset page: https://huggingface.co/datasets/simon3000/genshin-voice.Openthoughts_math_30k_opsdrobocasa365-pretrain-mg
Pretraining (MimicGen) — atomic
MimicGen-generated rollouts across 60 atomic tasks (~10,000 demos/task). 1,615 hours total, generated by scripted augmentation from human demonstrations.
Part of the RoboCasa365 collection. Flat LeRobot v3.0 mirror of RoboCasa365 — standard layout, drop-in loadable.
Stats
Episodes: 536,030
Frames: 116,246,439 (20 fps → 1615 h)
Tasks: 720 (natural-language phrasings; underlying RoboCasa task classes: 60)
Cameras: 3 × 256×256 h264… See the full description on the dataset page: https://huggingface.co/datasets/ember-lab-berkeley/robocasa365-pretrain-mg.3d-front-arjailbreak-deepseek-v3.2-expAnyWord-3MDataset from AnyText: Multilingual Visual Text Generation And Editing.
Dataset description from Anytext Team:
Currently, there is a relative scarcity of public datasets for text generation tasks, especially those involving non-Latin script languages. To address this, we introduce a large-scale multilingual dataset called AnyWord-3M. The images in this dataset are sourced from Noah-Wukong, LAION-400M, and OCR recognition datasets such as ArT, COCO-Text, RCTW, LSVT, MLT, MTWI, ReCTS, etc. These… See the full description on the dataset page: https://huggingface.co/datasets/stzhao/AnyWord-3M.coral-v3
CoRal: Danish Conversational and Read-aloud Dataset
Version 3.0
Dataset Overview
CoRal is a comprehensive Automatic Speech Recognition (ASR) dataset designed to capture the diversity of the Danish language across various dialects, accents, genders, and age groups. The primary goal of the CoRal dataset is to provide a robust resource for training and evaluating ASR models that can understand and transcribe spoken Danish in all its variations.
Key Features… See the full description on the dataset page: https://huggingface.co/datasets/CoRal-project/coral-v3.gs-images-v3HLE_200-3esb-datasets-earnings22-validation-tiny-filteredA filtered (<=30s duration) slice (512 samples) of the Earnings22 dataset.
def add_duration(sample):
y, sr = sample['audio']["array"], sample['audio']["sampling_rate"]
sample['duration_ms']=librosa.get_duration(y=y, sr=sr) * 1000
return sample
tedlium = load_dataset("esb/datasets", "earnings22", split='validation', trust_remote_code=True)
# compute duration to filter
tedlium = tedlium.map(add_duration)
tedlium = tedlium.select(range(512))
# Whisper max supported duration
tedlium… See the full description on the dataset page: https://huggingface.co/datasets/D4nt3/esb-datasets-earnings22-validation-tiny-filtered.TrainingData_Stage3
AnchorSR Stage3 · metric-v1.0
直接选择 Small / Large
配置
训练题数
用途
small
1,000,000
先验证答案监督／先验恢复，按新版 Large 联合分布抽样
large
89,801,853
筛选后的完整训练集合，包含 Small 全部样本
from datasets import load_dataset
data = load_dataset('AnchorSR/TrainingData_Stage3', 'small', # 或 large
revision='metric-v1.0', streaming=True)
这是对 scaling-v1.0 的语义筛选与统一任务分类，不是增加新数据源。
Large 从 89,828,269 题保留 89,801,853 题，隔离 26,416 题。
旧标签 scaling-v1.0 / video-v1.0 / large-v1.0… See the full description on the dataset page: https://huggingface.co/datasets/AnchorSR/TrainingData_Stage3.VidaForge-3M
3.14 million scene-level video clips with multi-level captions, camera labels, semantic tags, quality signals, and duplicate groups.
Paper
·
VidaForge Code
·
Project Blog
·
Source Dataset
Overview
VidaForge-3M is a large-scale video pretraining dataset produced with
VidaForge, an open data pipeline for
building and studying video foundation model pretraining data. The pipeline and
dataset are described in the paper
VidaForge: Open Research Infrastructure… See the full description on the dataset page: https://huggingface.co/datasets/VidaForge/VidaForge-3M.shrutilipiblip3-kale
🥬 BLIP3-KALE:Knowledge Augmented Large-scale Dense Captions
BLIP3-KALE is an open-source dataset of 218 million image-text pairs, featuring knowledge-augmented dense captions combining web-scale knowledge with detailed image descriptions.
Paper: [To be added]
Uses
BLIP3-KALE is designed to facilitate research in multimodal pretraining. The dataset can be used for training large multimodal models that require factually grounded, dense image captions. It has already been an… See the full description on the dataset page: https://huggingface.co/datasets/Salesforce/blip3-kale.paired-llama-3.2-1b-embeddings-lmsys-chat-1m
Paired Llama 3.2 1B Token Embeddings (LMSYS-Chat-1M)
This dataset contains paired activations corresponding to single token locations extracted from Meta's Llama 3.2 1B Instruct on conversations from LMSYS-Chat-1M.
Embeddings are provided for layers 5 through 14, which capture the most interesting intermediate representations.
This dataset was built to study things like:
Learning different basis for activations at a given layer
Studying if there are cases where position encodes… See the full description on the dataset page: https://huggingface.co/datasets/scaleinvariant/paired-llama-3.2-1b-embeddings-lmsys-chat-1m.CC-MAIN-2025-33zenless-voice
Zenless Voice
Zenless Voice is a dataset of voice lines from the popular game Zenless Zone Zero.
Hugging Face 🤗 Zenless-Voice
ModelScope Zenless-Voice
Per-speaker downloads are grouped by language and WAV count. Browse every archive in the ZIP index.
Last update at 2026-09-17, game version 3.2.0
406720 wavs
78785 without speaker (19%)
123429 without transcription (30%)
83509 without inGameFilename (21%)
Speaker archives contain 327,935 WAVs in 4,322 ZIPs. The 78,785 rows… See the full description on the dataset page: https://huggingface.co/datasets/simon3000/zenless-voice.