QASC is a question-answering dataset with a focus on sentence composition. It consists of 9,980 8-way multiple-choice questions about grade school science (8,134 train, 926 dev, 920 test), and comes with a corpus of 17M sentences. An example of 'validation' looks as follows. The data fields are the same among all splits. - id: a string feature. - question: a string feature. - choices: a dictionary feature containing: - text: a string feature. - label: a string feature. - answerKey: a string feature. - fact1: a string feature. - fact2: a string feature. - combinedfact: a string feature. - formattedquestion: a string feature. The dataset is released under CC BY 4.0 license. Thanks to…
SAVRN Model Hub
AI Training Datasets
Each dataset with its card, its structure (every split, row count and column type), its files, its license, and the models that disclose training on it.
2,760Models
859Datasets
254Papers
1,692Publishers
5,040Sourced relationships
Updated 2026-09-18 · How the library is built
859 datasets, sorted by most downloaded.
Training data for the LLaVA-OneVision-2 multimodal model family. The release contains large-scale video data at several duration ranges, video captions and source mappings, and spatial-reasoning data used for mid-training. The dataset is split across two Hugging Face repositories because of its size: If you need the long-video data, download the video shards from Part 2 and use the corresponding captions and mapping files from Part 1. The directory names in Part 2 are upload prefixes. Their duration mapping is: The mapping above was verified from the sample paths stored inside the WebDataset archives. For the ~180-second split: 1. Download video shards from tom/ in Part 2. 2. Read captions…
MegaWika is a multi- and crosslingual text dataset containing 30 million Wikipedia passages with their scraped and cleaned web citations. The passages span 50 Wikipedias in 50 languages, and the articles in which the passages were originally embedded are included for convenience. Where a Wikipedia passage is in a non-English language, an automated English translation is provided. Furthermore, nearly 130 million English question/answer pairs were extracted from the passages, and FrameNet events occurring in the passages are detected using the LOME FrameNet parser. The pipeline through which MegaWika was created is complex, and is described in more detail in the paper (linked above), but the…
The CT-RATE Team organizes the VLM3D Challenge VLM3D 2026 (2nd Edition) → Challenge Finals at MICCAI 2026 VLM3D 2025 (1st Edition) → Challenge Finals at MICCAI 2025 • Workshop at ICCV 2025 The CT-RATE Team is developing the MR-RATE Dataset A large-scale brain MRI dataset with paired radiology reports for training 3D Foundation Models from a Multimodal Dataset for 3D Computed Tomography
GPQA is a multiple-choice, Q&A dataset of very hard questions written and validated by experts in biology, physics, and chemistry. When attempting questions out of their own domain (e.g., a physicist answers a chemistry question), these experts get only 34% accuracy, despite spending >30m with full access to Google. We request that you do not reveal examples from this dataset in plain text or images online, to reduce the risk of leakage into foundation model training corpora. We present GPQA, a challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. We ensure that the questions are high-quality and extremely difficult: experts who…
The dataset is a collection of simple math word problems focused on arithmetics. It is derived from. The main addition in this dataset variant is the chain column. It was created by converting the solution to a simple html-like language that can be easily parsed (e.g. by BeautifulSoup). The data contains 3 types of tags: This variant of the dataset is intended for training Chain-of-Thought reasoning models able to use external tools to enhance the factuality of their responses. This dataset presents in-context scenarios where models can outsource the computations in the reasoning chain to a calculator. We provide 2 variants of the dataset. In the first one, the data splits correspond to the…
This dataset was created using LeRobot. One of the biggest open-source dataset for robotics with 27.044,326 frames, 92,223 episodes, 31,308 unique task description in natural language. Ported from Tensorflow Dataset format (2TB) to LeRobotDataset format (400GB) with the help from IPEC-COMMUNITY.
SWE-bench Verified is a subset of 500 samples from the SWE-bench test set, which have been human-validated for quality. SWE-bench is a dataset that tests systems’ ability to solve GitHub issues automatically. See this post for more details on the human-validation process. The dataset collects 500 test Issue-Pull Request pairs from popular Python repositories. Evaluation is performed by unit test verification using post-PR behavior as the reference solution. The original SWE-bench dataset was released as part of SWE-bench: Can Language Models Resolve Real-World GitHub Issues? Want to run inference now? This dataset only contains the problemstatement (i.e. issue text) and the basecommit which…
This dataset was created using LeRobot.
arXiv Papers by Subject A reorganised version of the nick007x/arxiv-papers dataset, partitioned by subject code, year, and month for efficient selective access. This dataset contains metadata for over 2.5 million arXiv papers, organised into a hierarchical directory structure that allows users to download only the specific subjects and time periods they need, rather than the entire dataset. Motivation The original
Alpaca is a dataset of 52,000 instructions and demonstrations generated by OpenAI's text-davinci-003 engine. This instruction data can be used to conduct instruction-tuning for language models and make the language model follow instruction better. The authors built on the data generation pipeline from Self-Instruct framework and made the following modifications: - The text-davinci-003 engine to generate the instruction data instead of davinci. - A new prompt was written that explicitly gave the requirement of instruction generation to text-davinci-003. - Much more aggressive batch decoding was used, i.e., generating 20 instructions at once, which significantly reduced the cost of data…
Egocentric-100K is the largest dataset of manual labor. You can visualize the dataset here. Egocentric-100K is state-of-the-art in hand visibility and active manipulation density compared to previous in-the-wild egocentric datasets. The complete 30,000 frame evaluation set is available at Egocentric-100K-Evaluation. Dataset Statistics Attribute Value Total Hours 100,405 Total Frames 10.8 billion Video Clips 2,010,759 Median Clip Length 180.0 seconds Mean
ACE-Data-0 Human-Centric Ambient Capture as Embodied Data Engine S-Lab, Nanyang Technological University, Singapore · ACE Robotics ACE turns real home environments into spatially calibrated, temporally synchronized recording studios for embodied AI. ▶ Demo video · Full story, figures, and interactive examples on the blog What this is Learning to act in the physical
This dataset contains BEHAVIOR-1K 2026 challenge demonstration trajectories in LeRobotDataset v3 format. The repository follows the LeRobotDataset v3 layout: - meta/info.json: dataset schema and path templates - meta/stats.json: feature statistics - meta/tasks.parquet: task metadata - meta/episodes/: per-episode metadata shards - data/: frame-level Parquet shards - videos/: encoded video shards The full dataset is large. For a quick smoke test, load a single episode without videos: To load the dataset at the published LeRobot v3 revision explicitly: The v3.0 Hub tag is intended to match the LeRobot v3.0 codebase-compatible dataset revision. This dataset is released under the MIT License.…
Model Collections
Hand-picked starting points, each with the reason it exists.
Collection · 4 entries
Embedding models for retrieval
Sentence and document embedding models used to build retrieval systems. Dimension and sequence length matter more than size here, and both come from the publisher.
Collection · 4 entries
Models that fit on one accelerator
Models whose publisher-reported parameter count puts them within reach of a single accelerator at common precisions. Memory needed depends on precision and serving configuration, so treat the parameter count as the starting point, not the answer.
Collection · 6 entries
Open-weight text models worth knowing
Widely used open-weight language models, chosen because each one is a distinct family rather than a variant of the one above it. Selection, not a ranking.
Collection · 3 entries
Speech and audio models
Recognition and synthesis models, grouped so the two directions are easy to compare.
Related SAVRN Research
The hub sits beside SAVRN's market data and infrastructure research: what models cost to run, and what it takes to run them.
SAVRN Index
What open models cost to run
The same open-weight model priced by every host that serves it, per million tokens.
Research Hub
Data center trackers and maps
Moratoriums, permits, power, water and capital behind the facilities that run these models.
Method
How the Model Hub is built
Sources, evidence labels, refresh behaviour, and the limits of every comparison here.