Independent project. Not a U.S. government website.Independent project. Not a United States government website.
Support our projectSupport usSupport our project · Buy me a coffee (opens Buy Me a Coffee)Conceptual Captions is a dataset consisting of ~3.3M images annotated with captions. In contrast with the curated style of other image caption annotations, Conceptual Ca…
112 likes11,530 downloads in the last 30 daysImage to textOther license
LightOnOCR-mix-0126 is a large-scale OCR training dataset built via distillation: a strong vision–language model is prompted to produce naturally ordered full-page trans…
112 likes1,426 downloads in the last 30 daysImage to textOther license
UltraInteract is a large-scale, high-quality alignment dataset specifically designed for complex reasoning tasks. For each instruction, it includes a preference tree con…
112 likes1,553 downloads in the last 30 daysMIT
Spark-234K: Skeleton-Guided Scientific Reasoning from Large-Scale Literature 🎉 Accepted to EMNLP 2026 Findings! Spark-234K is a scientific reasoning dataset containing…
112 likes3,732 downloads in the last 30 daysText generationApache 2.0
112 likes547 downloads in the last 30 daysNo license given
AudioSet is a dataset of 10-second clips from YouTube, annotated into one or more sound categories, following the AudioSet ontology. Supported Tasks and Leaderboards aud…
111 likes48,379 downloads in the last 30 daysAudio classificationCC BY 4.0
The Tome is a curated dataset designed for training large language models with a focus on instruction following. It was used in the training of our Arcee-Nova/Spark mode…
111 likes303 downloads in the last 30 daysMIT
和谐历史档案馆数据集 - Banned Historical Archives Datasets 和谐历史档案馆数据集包含已录入 和暂未未录入的原始文件。 目录结构 banned-historical-archives.github.io # 已录入该网站的原始数据，不定期从 github 仓库中同步 raw # 原始文件 config…
111 likes1,101,900 downloads in the last 30 daysNo license given
"instruction": "将以下三个句子组合成一个有意义的段落。\n狗是人类最好的朋友。它们非常聪明，可以进行各种活动。如果你喜欢散步，狗可以成为你一起散步的伙伴。", "input": "", "output": "狗是人类最好的朋友，它们非常聪明，可以进行各种活动。如果你喜欢散步，狗可以成为你一起散步的伙伴。出门散步是一种良好…
111 likes547 downloads in the last 30 daysGPL 3.0
DramaBench: Drama Script Continuation Dataset Dataset Summary DramaBench is a comprehensive benchmark dataset for evaluating drama script continuation capabilities of la…
111 likes272 downloads in the last 30 daysText generationMIT
EvasionBench is a benchmark dataset for detecting evasive answers in earnings call Q&A sessions. The task is to classify how directly corporate management addresses ques…
111 likes435 downloads in the last 30 daysText classificationApache 2.0
This dataset contains submissions from a prompt hacking competition. An in-depth analysis of the dataset has been accepted at the EMNLP 2023 conference. 📊👾 Submissions…
111 likes1,183 downloads in the last 30 daysMITGated
To apply eyeshadow without a brush, should I use a cotton swab or a toothpick? Questions requiring this kind of physical commonsense pose a challenge to state-of-the-art…
111 likes85,132 downloads in the last 30 daysQuestion answeringUnknown license
english_quotes is a dataset of all the quotes retrieved from goodreads quotes. This dataset can be used for multi-label text classification and text generation. The cont…
110 likes2,974 downloads in the last 30 daysText classificationNo license given
Final self-alignment training dataset for StarCoder2-Instruct. seed: Contains the seed Python function concepts: Contains the concepts generated from the seed instructio…
110 likes32,881 downloads in the last 30 daysODC-By
Project Web: Arxiv Technical Report: Codes: Abstract Click Here High-quality instruction data is critical for aligning large language models (LLMs). Although some models…
110 likes601 downloads in the last 30 daysLlama 3.3 license
DRAFT LIST OF ALL HUMAN DISEASES Way more data coming. License is AGPL to avoid predatory players, I don't care if you or your startup use it. If you see an issue commen…
110 likes213 downloads in the last 30 daysAGPL 3.0
AgentWorldBench is a comprehensive evaluation benchmark for language world models, constructed from real-world observations of frontier model trajectories on established…
110 likes1,055 downloads in the last 30 daysText generationApache 2.0
Open-Jev turns a state and a question into a typed decision: a yes/no probability, a distribution over choices, independent label probabilities, or a discrete numeric/or…
110 likes7,532 downloads in the last 30 daysText classificationCC0 1.0
This data set contains over 6,000 medical terms and their wikipedia text. It is intended to be used on a downstream task that requires medical terms and their wikipedia…
109 likes427 downloads in the last 30 daysText classificationGPL 3.0
SynLogic is a comprehensive synthetic logical reasoning dataset designed to enhance logical reasoning capabilities in Large Language Models (LLMs) through reinforcement…
109 likes1,214 downloads in the last 30 daysText generationMIT
Project page | Paper | Code Dataset Description AudioSkills-XL is a large-scale audio question-answering (AQA) dataset designed to develop (large) audio-language models…
109 likes1,556 downloads in the last 30 daysAudio text to textOther license
Midjourney is an independent research lab whose broad mission is to "explore new mediums of thought". In 2022, they launched a text-to-image service that, given a natura…
109 likes269 downloads in the last 30 daysApache 2.0
This dataset was generated using teich by TeichAI Prepare these datasets for supervised fine-tuning in just a few lines of code — see the Conversion section below. DeepS…
109 likes3,911 downloads in the last 30 daysText generationNo license given
This is raw, semi-structured data — not a ready-to-train dataset. It still requires further cleaning and preparation (deduplication, severity/label normalization, filter…
109 likes1,115 downloads in the last 30 daysText classificationOther license
The data includes examples that might be disturbing, harmful or upsetting. It includes a range of harmful topics such as discriminatory language and discussions about ab…
108 likes8,501 downloads in the last 30 daysText classificationODC-ByGated
DatBench: Discriminative, Faithful, and Efficient VLM Evaluations DatBench is a curated evaluation suite for vision–language models (VLMs) designed to be faithful, discr…
108 likes2,623 downloads in the last 30 daysNo license given
FINAL Bench: Functional Metacognitive Reasoning Benchmark "Not how much AI knows — but whether it knows what it doesn't know, and can fix it." --- Overview FINAL Bench (…
108 likes394 downloads in the last 30 daysText generationApache 2.0
This dataset consists of 52k samples generated through Glaive for the task of function calling, in the following format- SYSTEM: You are an helpful assistant who has acc…
108 likes726 downloads in the last 30 daysText generationApache 2.0
PRISM is a diverse human feedback dataset for preference and value alignment in Large Language Models (LLMs). It maps the characteristics and stated preferences of human…
108 likes4,131 downloads in the last 30 daysCC
What is Megalith-10m? Megalith-10m is a dataset of ~10 million links to Flickr images that were categorized as "photo" with license info of: No known copyright restricti…
108 likes907 downloads in the last 30 daysMIT
High-Quality Traditional Chinese Medicine Dataset from Non-Internet Sources - SFT/IFT 该数据集经过大量人力和资源的投入精心构建，以共建LLM高质量中文社区为己任。 包含约1GB的中医各个领域临床案例、名家典籍、医学百科，名词解释等优质问答内容，涵盖全面…
108 likes792 downloads in the last 30 daysTable question answeringApache 2.0
The LLaVA-CoT-100k dataset is introduced in the paper LLaVA-CoT: Let Vision Language Models Reason Step-by-Step. This dataset is designed to enable Vision-Language Model…
108 likes2,384 downloads in the last 30 daysVisual question answeringApache 2.0
注意：此数据集是由ChatGPT产生的，未经过严格校验，题目或解题过程可能包含错误。使用过程中请注意这一点。 样例 { "instruction": "题目：小华手里有一个装满糖果的袋子，共有12个，小明想知道里面有多少颗糖果，于是他问小华：“你手里的糖果袋子里有偶数个糖果吗？”小华回答：“有，而且多于10颗。”请问小华手里的糖果袋子里…
107 likes772 downloads in the last 30 daysGPL 3.0
[2025/02/18]: We add the original captions of PubMedVision in PubMedVision_Original_Caption.json, as well as the Chinese version of PubMedVision in PubMedVision_Chinese.…
107 likes2,017 downloads in the last 30 daysQuestion answeringApache 2.0
Glaive-code-assistant is a dataset of ~140k code problems and solutions generated using Glaive’s synthetic data generation platform. The data is intended to be used to m…
107 likes1,118 downloads in the last 30 daysApache 2.0
Agent-FLAN: Designing Data and Methods of Effective Agent Tuning for Large Language Models This page holds the dataset proposed in Agent-FLAN, which consists of AgentIns…
107 likes805 downloads in the last 30 daysApache 2.0
The Open Australian Legal Corpus by Isaacus, a foundational legal AI research company, is the first and only multijurisdictional open corpus of Australian legislative an…
107 likes17,651 downloads in the last 30 daysText generationOpen australian legal corpus
107 likes282 downloads in the last 30 daysCC BY 2.0
The OpenCoder dataset is composed of the following datasets: opc-sft-stage1: the sft data used for opencoder sft-stage1 opc-sft-stage2: the sft data used for opencoder s…
107 likes25,444 downloads in the last 30 daysMIT
This corpus is an attempt to recreate the dataset used for training XLM-R. This corpus comprises of monolingual data for 100+ languages and also includes data for romani…
107 likes1,170 downloads in the last 30 daysText generationUnknown license
Tiny ImageNet contains 100000 images of 200 classes (500 for each class) downsized to 64×64 colored images. Each class has 500 training images, 50 validation images, and…
107 likes27,288 downloads in the last 30 daysImage classificationNo license given
📝 Blog post: 💻 GitHub: IFStruct is a benchmark for structured-output compliance: can a model produce valid JSON/YAML that follows a requested schema, when the requirem…
106 likes756 downloads in the last 30 daysText generationApache 2.0
4,000 examples of argilla/distilabel-capybara-dpo-7k-binarized with chosen score =4. 3,000 examples of argilla/distilabel-intel-orca-dpo-pairs with chosen score =8. 3,00…
106 likes339 downloads in the last 30 daysText generationApache 2.0
🚀 LLaVA-One-Vision-1.5-Mid-Training-85M Dataset is being uploaded 🚀 Upload Status All Completed: ImageNet-21k、LAIONCN、DataComp-1B、Zero250M、COYO700M、SA-1B、MINT、Obelics…
106 likes567,180 downloads in the last 30 daysApache 2.0
OpenThoughts-Agent is an open-source effort to curate the best datasets for training agents. Our first release includes datasets, models and our research codebase. OpenT…
106 likes2,464 downloads in the last 30 daysApache 2.0
ODA-Math-460k is a large-scale math reasoning dataset curated from top-performing open mathematics corpora (selected via the OpenDataArena leaderboard) and refined throu…
106 likes467 downloads in the last 30 daysQuestion answeringCC BY-NC 4.0
Paper: Data Loading from datasets import load_dataset dataset = load_dataset("peiyi9979/Math-Shepherd") Data Instance Every instance consists of three data fields: "inpu…
106 likes2,268 downloads in the last 30 daysNo license given
A Large-Scale Multi-Disciplinary Scientific Multimodal Dataset S1-MMAlign is a large-scale, multi-disciplinary multimodal dataset comprising over 15.5 million high-quali…
106 likes2,031 downloads in the last 30 daysImage to textCC BY-NC 4.0
Optimized CAD assets with configured collision and physical properties. Manually reviewed scenes Physics configuration reviewed for every household scene. Scalable task…
106 likes17,926 downloads in the last 30 daysRoboticsCC BY-NC-SA 4.0Gated
DOLMino dataset mix for OLMo2 stage 2 annealing training. Mixture of high-quality data used for the second stage of OLMo2 training. Source Sizes Name Category Tokens Byt…
105 likes12,620 downloads in the last 30 daysText generationODC-By
Cleaned, section-chunked training corpus for a small bilingual LLM (Arabic + English), combining a curated Egyptian-history collection with general Wikipedia coverage fr…
105 likes3,136 downloads in the last 30 daysText generationCC BY-SA 4.0
xP3x (Crosslingual Public Pool of Prompts eXtended) is a collection of prompts & datasets across 277 languages & 16 NLP tasks. It contains all of xP3 + much more! It is…
105 likes37,223 downloads in the last 30 daysOtherApache 2.0
VQA-RAD is a dataset of question-answer pairs on radiology images. The dataset is intended to be used for training and testing Medical Visual Question Answering (VQA) sy…
105 likes8,905 downloads in the last 30 daysVisual question answeringCC0 1.0
GPT Role-play Realm Dataset: The AI-generated character compendium This is a dataset of GPT-generated characters made to increase the ability of open-source language mod…
105 likes213 downloads in the last 30 daysText generationCC BY 4.0
StackMathQA: A Curated Collection of 2 Million Mathematical Questions and Answers Sourced from Stack Exchange StackMathQA is a meticulously curated collection of 2 milli…
105 likes1,691 downloads in the last 30 daysText generationCC BY 4.0
Abordagem de IA composta para geração de personas baseada em distribuições do mundo real Visão Geral do Conjunto de Dados (Dataset Overview): Nemotron-Personas-Brazil é…
105 likes1,033 downloads in the last 30 daysText generationCC BY 4.0
105 likes3,265 downloads in the last 30 daysText generationApache 2.0
SAGE-3D Collision Mesh: Physics-Enabled Collision Bodies for 3D Gaussian Scenes Paper | Project Page | Code High-precision collision geometry dataset extracted from 1,00…
105 likes1,542 downloads in the last 30 daysRoboticsCC BY-NC 4.0
感谢 Open-Orca/OpenOrca 数据集的发布，给广大NLP研究人员和开发者带来了宝贵的资源！ 这是一个对 Open-Orca/OpenOrca 数据集中文翻译的版本，翻译引擎为 Google 翻译，希望能给中文 LLM 研究做出一点点贡献。 Dataset Summary The OpenOrca dataset is a…
105 likes156 downloads in the last 30 daysText classificationMIT
Anthropic's HH dataset reformatted into prompt, chosen, rejected samples.
104 likes421 downloads in the last 30 daysNo license given
104 likes2,313 downloads in the last 30 daysMIT
We propose KMMLU, a new Korean benchmark with 35,030 expert-level multiple-choice questions across 45 subjects ranging from humanities to STEM. Unlike previous Korean be…
104 likes12,905 downloads in the last 30 daysMultiple choiceCC BY ND 4.0
AceReason-1.1-SFT is a diverse and high-quality supervised fine-tuning (SFT) dataset focused on math and code reasoning. It serves as the SFT training data for AceReason…
104 likes2,750 downloads in the last 30 daysText generationCC BY 4.0
I generated a dataset to test the knowledge scope and capabilities of MiniMax H3. Samples are of various aspect sizes, and cover a wide range of media types and themes.…
104 likes3,040 downloads in the last 30 daysNo license given
We provide an Amazon product reviews dataset for multilingual text classification. The dataset contains reviews in English, Japanese, German, French, Chinese and Spanish…
103 likes1,756 downloads in the last 30 daysSummarizationOther license
ELYZA-tasks-100: 日本語instructionモデル評価データセット Data Description 本データセットはinstruction-tuningを行ったモデルの評価用データセットです。詳細は リリースのnote記事 を参照してください。 特徴: 複雑な指示・タスクを含む100件の日本語データです。 役に立つA…
103 likes1,593 downloads in the last 30 daysCC BY-SA 4.0
In collaboration with Ravel Law, Harvard Law Library digitized over 40 million U.S. court decisions consisting of 6.7 million cases from the last 360 years into a datase…
103 likes402 downloads in the last 30 daysText generationCC0 1.0Gated
350B+ of highly educational tokens from PDFs 📄 What is it? 📚 FinePDFs-Edu dataset consists of 350B+ tokens of educational PDFs filtered from 📄 FinePDFs dataset coveri…
103 likes13,209 downloads in the last 30 daysText generationODC-By
This dataset contains 80,036 trajectories generated by a software engineering agent based on the SWE-agent framework, using various models as action generators. In these…
103 likes12,541 downloads in the last 30 daysCC BY 4.0
👩💻 Authors: Rabeeh Karimi Mahabadi, Sanjeev Satheesh 📘 Paper: Nemotron-cc-math: A 133 Billion-Token-Scale High Quality Math Pretraining Dataset 📝 Blog: Nemotron-cc-…
103 likes8,574 downloads in the last 30 daysText generationOther licenseGated
📮 [2025-08] We have updated to a larger dataset, which contains nearly 80 million samples, with significant improvements in both data quality and diversity. [link] 📮 […
103 likes453 downloads in the last 30 daysImage to textOpenRAIL
The Large-scale Open Time Series Archive (LOTSA) is a collection of open time series datasets for time series forecasting. It was collected for the purpose of pre-traini…
103 likes81,123 downloads in the last 30 daysApache 2.0
103 likes4,926 downloads in the last 30 daysNo license given
This dataset is a high-quality reasoning corpus distilled from DeepSeek-R1-0528, an improved version of the DeepSeek-R1 large language model. Compared to its initial rel…
102 likes843 downloads in the last 30 daysText generationNo license given
MASSIVE is a parallel dataset of 1M utterances across 51 languages with annotations for the Natural Language Understanding tasks of intent prediction and slot annotation…
102 likes15,290 downloads in the last 30 daysText classificationCC BY 4.0
ShareGPT-4o-Image is a large-scale and high-quality image generation dataset, where all images are produced by GPT-4o’s image generation capabilities. This dataset is de…
102 likes919 downloads in the last 30 daysText to imageApache 2.0
Original dataset introduced by Jin et al. in What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Exams Citation inform…
102 likes43,251 downloads in the last 30 daysCC BY 4.0
You need to first download the data from here and then apply for access to the original Laion-face dataset by completing the required agreement (github). Once approved,…
102 likes215 downloads in the last 30 daysImage to textCC BY 4.0
More Information needed
102 likes233 downloads in the last 30 daysNo license given
Paper: CoEdIT: Text Editing by Task-Specific Instruction Tuning Authors: Vipul Raheja, Dhruv Kumar, Ryan Koo, Dongyeop Kang Project Repo: Dataset Summary This is the dat…
101 likes1,954 downloads in the last 30 daysText generationApache 2.0
[ICLR 2026] OmniWorld: A Multi-Domain and Multi-Modal Dataset for 4D World Modeling 🎉NEWS [2026.3.21] 🔥 OmniWorld-Game with Metric Scale is now released! Check out our…
101 likes58,456 downloads in the last 30 daysText to videoCC BY-NC-SA 4.0
A dataset of real-world data analytics job postings from 2023, collected and processed by Luke Barousse. Background I've been collecting data on data job postings since…
101 likes7,071 downloads in the last 30 daysApache 2.0
101 likes298 downloads in the last 30 daysApache 2.0
This is a filtered and metadata enriched version of open-thoughts/OpenThoughts-114k. While the original dataset is a valuable resource containing DeepSeek-R1 outputs, it…
101 likes5,678 downloads in the last 30 daysNo license given
We present InternVid-10M-FLT, a subset of this dataset, consisting of 10 million video clips, with generated high-quality captions for publicly available web videos. Dow…
101 likes199 downloads in the last 30 daysFeature extraction (embeddings)CC BY-NC-SA 4.0Gated
The ultimate medical reasoning dataset - combining 7 state-of-the-art AI models with fair distribution deduplication. 1.79 million unique samples with 3.78 billion token…
101 likes2,050 downloads in the last 30 daysText generationApache 2.0
Quick Start · Overview · Tasks · Comparison · Dataset Stats · Data Sources · Pipeline · Schema · License · Citation This is the official MathNet v0. A larger version v1…
101 likes85,975 downloads in the last 30 daysQuestion answeringCC BY 4.0
This is an unoffical reupload of Code_bagel. You can find the original dataset here: A coding bagel, with everything coding related Around 800 million tokens of unique c…
101 likes152 downloads in the last 30 daysApache 2.0
Zyda-2 is a 5 trillion token language modeling dataset created by collecting open and high quality datasets and combining them and cross-deduplication and model-based qu…
101 likes32,401 downloads in the last 30 daysText generationODC-By
100 likes14,076 downloads in the last 30 daysApache 2.0
A small subset (~0.1%) of the-stack dataset, each programming language has 10,000 random samples from the original dataset. The dataset has 2.6GB of text (code). Languag…
100 likes23,779 downloads in the last 30 daysText generationNo license givenGated
Thanks for Guanaco Dataset and Belle Dataset This dataset was created by merging the above two datasets in a certain format so that they can be used for training our cod…
100 likes105 downloads in the last 30 daysGPL 3.0
The AMI Meeting Corpus consists of 100 hours of meeting recordings. The recordings use a range of signals synchronized to a common timeline. These include close-talking…
100 likes19,863 downloads in the last 30 daysSpeech recognitionCC BY 4.0
This is a repository containing a subset of the DISC-Med-SFT Dataset. Check DISC-MedLLM for more information.
100 likes255 downloads in the last 30 daysQuestion answeringApache 2.0
100 likes3,144 downloads in the last 30 daysApache 2.0
Code problem-solving data with reasoning, direct-answer, and task-synthesis subsets. This repository is part of the K2 Horizon collection. The repository is organized in…
100 likes24,275 downloads in the last 30 daysText generationApache 2.0
100 likes262 downloads in the last 30 daysNo license given
Multimodal-Mind2Web is the multimodal version of Mind2Web, a dataset for developing and evaluating generalist agents for the web that can follow language instructions to…
100 likes11,191 downloads in the last 30 daysOpenRAIL
This dataset contains four multi-view image sequences used in our paper "TaoAvatar: Real-Time Lifelike Full-Body Talking Avatars for Augmented Reality via 3D Gaussian Sp…
100 likes462 downloads in the last 30 daysCC BY-NC 4.0Gated
Help keep USASI useful.
Optional. No USASI account required. Payment takes place on the linked provider’s website (Buy Me a Coffee).
About supporting this project