datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
s2orc_full
S2ORC Full — Semantic Scholar Open Research Corpus
A complete redistribution of the S2ORC dataset in Parquet format on Hugging Face, containing 14.5 million academic papers with full text, structured metadata, and citation information.
Dataset Description
S2ORC (Semantic Scholar Open Research Corpus) is a general-purpose corpus for NLP and text mining research over scientific papers, originally developed by the Allen Institute for AI. This version provides the full… See the full description on the dataset page: https://huggingface.co/datasets/AlgorithmicResearchGroup/s2orc_full.migration-bench-java-full
MigrationBench
1. 📖 Overview
🤗 MigrationBench
is a large-scale code migration benchmark dataset at the repository level,
across multiple programming languages.
Current and initial release includes java 8 repositories with the maven build system… See the full description on the dataset page: https://huggingface.co/datasets/AmazonScience/migration-bench-java-full.MMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking
MMFineReason-Full-2.3M
The Complete Pre-Selection Dataset — Before Quality Filtering
📖 Overview
MMFineReason-Full-2.3M is the complete pre-selection dataset containing 2.3M samples and 8.8B solution tokens, generated through our reasoning distillation pipeline before the data selection stage. This dataset includes all samples that passed basic template and length validation, but have not undergone correctness verification filtering.
🎯 Key Characteristics… See the full description on the dataset page: https://huggingface.co/datasets/OpenDataArena/MMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking.s2orc_full
S2ORC Full — Semantic Scholar Open Research Corpus
A complete redistribution of the S2ORC dataset in Parquet format on Hugging Face, containing 14.5 million academic papers with full text, structured metadata, and citation information.
Dataset Description
S2ORC (Semantic Scholar Open Research Corpus) is a general-purpose corpus for NLP and text mining research over scientific papers, originally developed by the Allen Institute for AI. This version provides the full… See the full description on the dataset page: https://huggingface.co/datasets/jedibear/s2orc_full.MMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking
MMFineReason-Full-2.3M
The Complete Pre-Selection Dataset — Before Quality Filtering
📖 Overview
MMFineReason-Full-2.3M is the complete pre-selection dataset containing 2.3M samples and 8.8B solution tokens, generated through our reasoning distillation pipeline before the data selection stage. This dataset includes all samples that passed basic template and length validation, but have not undergone correctness verification filtering.
🎯 Key Characteristics… See the full description on the dataset page: https://huggingface.co/datasets/ericktwo/MMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking.arxiv-tex-corpus-fullarxiv-tex-corpus-full (80GB)
Large-scale LaTeX corpus from arXiv (math, CS, physics, statistics)
📄 Paper: https://arxiv.org/abs/2602.17288
📚 Overview
arxiv-tex-corpus-full (80GB) is a large-scale dataset of LaTeX source content extracted from papers hosted on arXiv.
This version contains approximately 80GB of structured JSONL data, restricted to the following arXiv categories:
math
cs
hep-th
hep-ph
quant-ph
stat.ML
stat.TH
The dataset is designed for research in:
Large… See the full description on the dataset page: https://huggingface.co/datasets/KiteFishAI/arxiv-tex-corpus-full.fulg
❄️FuLG
The FuLG dataset is a comprehensive Romanian language corpus comprising 150 billion tokens, carefully
extracted from Common Crawl. This extensive dataset is the result of rigorous filtering and deduplication
processes applied to 95 Common Crawl snapshots. The compressed dataset has 289 GB.
For more details, check the arXiv preprint.
How do I download this?
Using 🤗 Datasets
from datasets import load_dataset
# Full dataset
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/faur-ai/fulg.WildChat-4.8M-Full
Dataset Card for WildChat-4.8M-Full
Dataset Description
Interactive Search Tool: https://wildvisualizer.com
WildChat paper: https://arxiv.org/abs/2405.01470
WildVis paper: https://arxiv.org/abs/2409.03753
Point of Contact: Yuntian Deng
Dataset Summary
WildChat-4.8M-Full is a collection of 4,743,336 conversations (out of 4,804,190 originally, after removing all conversations flagged with "sexual/minors" by OpenAI Moderation) between human users and… See the full description on the dataset page: https://huggingface.co/datasets/yuntian-deng/WildChat-4.8M-Full.DeepSWEGym2-Full
Dataset Description
This dataset is a merge containing many high quality SWE datasets, it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is NOT specifically filtered for rows with complex/long code problems in the original datasets, despite still having an average row size of 169.08kb, a total uncompressed size of 19.80GB, and a total of 122791 examples.
Dataset Details
Curated by: MoreThought
Funded by:… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym2-Full.R2E-Gym-Full
R2E-Gym Subset Filtered for MAGRPO
Filtered subset of R2E-Gym optimized for 2-agent MAGRPO training with 7B models.
Dataset Statistics
Total instances: 167
Format: Issue description + Oracle files in prompt
Optimized for: 2-agent collaboration, 7B models
Filtering Criteria (SWE-bench Lite Style)
Problem statement: >40 words (up to 500 for context window)
Must have non-empty oracle patch (non-test file changes)
File count: Exactly 1 oracle file (single-file… See the full description on the dataset page: https://huggingface.co/datasets/ryankamiri/R2E-Gym-Full.the-stack-v2-train-full-ids
The Stack v2
The dataset consists of 4 versions:
bigcode/the-stack-v2: the full "The Stack v2" dataset
bigcode/the-stack-v2-dedup: based on the bigcode/the-stack-v2 but further near-deduplicated
bigcode/the-stack-v2-train-full-ids: based on the bigcode/the-stack-v2-dedup dataset but further filtered with heuristics and spanning 600+ programming languages. The data is grouped into repositories. <-- you are here
bigcode/the-stack-v2-train-smol-ids: based on the… See the full description on the dataset page: https://huggingface.co/datasets/bigcode/the-stack-v2-train-full-ids.DeepSWEGym-Full
Dataset Description
This dataset is a merged version of all the SWE-bench/SWE-smith-lang datasets (88k rows total), it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is NOT specifically filtered for rows with complex/long code problems in the original datasets, despite still having an average row size of 85.4kb, a total uncompressed size of 7.53GB, and 88130 examples total.
Dataset Details
Curated by:… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym-Full.turk-ictihat-kararlari-fulltext
Türk İçtihat Kararları - Tam Metin (UYAP e-Bedesten)
Adalet Bakanlığı UYAP Mevzuat ve İçtihat Programı (bedesten.adalet.gov.tr)
üzerinden derlenen Yargıtay kararları veri seti. Kavram bazlı arama
sorguları ile toplanmıştır.
Boyut
Kayıt: 9,899,589 benzersiz karar
Terim dağılımı: {"bosanma": 100052, "karar": 9894542, "nafaka": 66241, "tazminat": 100052}
Yıl aralığı: 1993-2026
Tam metni gelen karar: 39,301 (artımlı olarak dolduruluyor)
Şema… See the full description on the dataset page: https://huggingface.co/datasets/muhammedturan/turk-ictihat-kararlari-fulltext.wikipedia-full
Dataset Card for Wikimedia Wikipedia
Dataset Summary
Wikipedia dataset containing cleaned articles of all languages.
The dataset is built from the Wikipedia dumps (https://dumps.wikimedia.org/)
with one subset per language, each containing a single train split.
Each example contains the content of one full Wikipedia article with cleaning to strip
markdown and unwanted sections (references, etc.).
All language subsets have already been processed for recent dump, and you… See the full description on the dataset page: https://huggingface.co/datasets/akahana/wikipedia-full.conseil-administratives-appel-full-documents
Décisions de Justice Administrative Françaises
Description du dataset
Ce dataset contient un corpus de décisions de justice administrative françaises extraites de la plateforme officielle Open Data de la Justice Administrative. Les documents sont publiés par le Conseil d'État dans le cadre de la politique d'ouverture des données publiques de la justice française. Les décisions sont diffusées au format XML et anonymisées conformément aux exigences légales relatives… See the full description on the dataset page: https://huggingface.co/datasets/hulk10/conseil-administratives-appel-full-documents.nyu-aco-ocr-full
nyu-aco-ocr-full
Scanned Arabic books from the Arabic Collections Online (ACO) archive, OCR'd page by page with the dots.ocr vision-language model served through vLLM. The archive spans seven partner collections: NYU, Princeton, Cornell, Columbia, AUB, AUC, and UAE National Archives.
Each PDF page is rendered at 200 DPI, OCR'd individually, and the pages of a book are joined into one markdown document separated by \n\n---\n\n. Each row is one complete book.
Schema… See the full description on the dataset page: https://huggingface.co/datasets/AdaMLLab/nyu-aco-ocr-full.conseil-detat-full-documents
Décisions du Conseil d'État (France)
Description
Ce dataset contient un corpus de décisions rendues par le Conseil d'État français, la plus haute juridiction de l'ordre administratif. Les décisions proviennent de la plateforme officielle Open Data de la Justice Administrative et sont diffusées au format XML anonymisé.
Le corpus rassemble les textes intégraux des décisions ainsi que plusieurs métadonnées permettant leur identification et leur traçabilité.
Source… See the full description on the dataset page: https://huggingface.co/datasets/hulk10/conseil-detat-full-documents.Arabic-VLM-Full-Pearl
💎 The Arabic VLM Dataset (Full Pearl Edition)
This repository contains the full, unreviewed dataset comprising 309K multimodal examples. This data was generated automatically using the agentic pipeline developed for the Pearl project, as described in our paper.
Disclaimer: This is the raw, synthetic data that has not been subject to human review. It was generated as part of the data creation process and is released for research purposes. It may contain noise, errors, or… See the full description on the dataset page: https://huggingface.co/datasets/MohamedRashad/Arabic-VLM-Full-Pearl.arXiv-full-text-chunked
Dataset Card for Dataset Name
This dataset card aims to be a base template for new datasets.
Dataset Details
Dataset Description
Curated by: [More Information Needed]
Funded by [optional]: [More Information Needed]
Shared by [optional]: [More Information Needed]
Language(s) (NLP): [More Information Needed]
License: [More Information Needed]
Dataset Sources [optional]
Repository: [More Information Needed]
Paper [optional]:… See the full description on the dataset page: https://huggingface.co/datasets/amrachraf/arXiv-full-text-chunked.llama-nemotron-science-reasoning-on-canonical-think-full
Llama-Nemotron science reasoning — Delphi canonical-think (COMPLETE, no length filter)
The complete reasoning:on science split of
nvidia/Llama-Nemotron-Post-Training-Dataset, converted once into the canonical
Delphi chat-template thinking format. 708,920 rows.
Unlike the cold-start warmup slice
open-athena/llama-nemotron-science-reasoning-on-le3000tok-100k
(and its -canonical-think variant), this build applies no length cap and no subsample — every
long-CoT science example is… See the full description on the dataset page: https://huggingface.co/datasets/laion/llama-nemotron-science-reasoning-on-canonical-think-full.flan2021-full
Task Name
FLAN-2021 -> 70
{
"ag_news_subset": 108497,
"ai2_arc/ARC-Challenge": 829,
"ai2_arc/ARC-Easy": 1927,
"aeslc": 13187,
"anli/r1": 15361,
"anli/r2": 41133,
"anli/r3": 91048,
"bool_q": 8343,
"cnn_dailymail": 259607,
"coqa": 6456,
"cosmos_qa": 22996,
"definite_pronoun_resolution": 1079,
"drop": 70045,
"fix_punct": 25690,
"gem/common_gen": 60936,
"gem/dart": 56724,
"gem/e2e_nlg": 30337,
"gem/web_nlg_en": 31899… See the full description on the dataset page: https://huggingface.co/datasets/aslawliet/flan2021-full.fully-open-meditron
Fully Open Meditron Corpus
👋 Join our LiGHT community.
📖 Check out the MeditronFO blog and MeditronFO preprint.
🔜 If you are a clinician join the MOOVE initiative here.
[Hugging Face]
[Preprint]
[GitHub]
[Dataset]
License: Apache 2.0 | Authors: LiGHT
[!Note]
A clinician-vetted training corpus for medical large language models, accompanying the paper Fully Open Meditron: An Auditable Pipeline for Clinical LLMs.
The… See the full description on the dataset page: https://huggingface.co/datasets/EPFLiGHT/fully-open-meditron.SWEUniverse-Repaired-Indist-full-not-SWE-bench-pro-matched
VmaxRL/SWEUniverse-Repaired-Indist-full-not-SWE-bench-pro-matched
This dataset contains a 350-row subset selected from the Indist SWEUniverse training rows.
Selection policy: three-way repo overlap with Bugpilot and LM-Modify, deduped by repo plus introduction patch, then balanced round-robin across overlapping repos.
Rows: 350
Selected repos: 19
Deduped overlap capacity: 468
Source dataset: VmaxRL/SWEUniverse-Repaired-Indist-full-not-SWE-bench-pro-matched
codeforces-editorial-full-2026-07-16
Codeforces Editorial Full - 2026-07-16
A deterministic Plan-CRL-compatible materialization of open-r1/codeforces, pinned to revision fbe3f6e903ee854eec2e69e9d96d0306cde59baf.
Size
train: 9556 problems
test: 468 problems
total: 10024 unique problems
safe fixed-output Plan-CRL evaluation rows: 1479
safe fixed-output rows with an editorial: 646
The upstream snapshot contains 10,024 unique problems, not 11,000+. The larger counts sometimes quoted for this corpus… See the full description on the dataset page: https://huggingface.co/datasets/1337xyz1337xyz/codeforces-editorial-full-2026-07-16.full-structured-instruction-sft-dataset
Full Structured + Instruction SFT Corpus
Unified SFT training corpus built from Glaive, Hermes, UltraChat, and synthetic structured-output data.
Dataset repo
mdonigian/full-structured-instruction-sft-datasetRelease date: 2026-03-11
Included files
train_full_sft.jsonl: full merged and shuffled SFT dataset
source_glaive.jsonl: processed Glaive subset
source_hermes.jsonl: processed Hermes subset
source_ultrachat.jsonl: processed UltraChat subset… See the full description on the dataset page: https://huggingface.co/datasets/mdonigian/full-structured-instruction-sft-dataset.math_full_minus_math500
MATH (minus MATH-500)
This dataset is derived from the original MATH dataset by Hendrycks et al.
(qwedsacf/competition_math) with all problems from the MATH-500 benchmark set removed.
Construction
Source: 12,500 problems from the MATH dataset by Hendrycks et al. (qwedsacf/competition_math)
Benchmark held out: 500 problems from the MATH-500 dataset (HuggingFaceH4/MATH-500)
Matching criterion: exact match on the problem field (see… See the full description on the dataset page: https://huggingface.co/datasets/rasbt/math_full_minus_math500.unpredictable_fullThe UnpredicTable dataset consists of web tables formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance. For more details please see the accompanying dataset card.flutter-full-examples-v1
Flutter Codegen: Full Examples
Synthetic dataset of complete Flutter/Dart widgets, each paired with the goal
that describes them and (optionally) starting code. Unlike flutter-codegen-diff-steps,
there's no step history or diff structure here -- each row is a single, standalone
goal -> complete file example.
This is the whole-code counterpart to flutter-diff-steps-v1, intended for
training/evaluating a baseline that generates the entire file in one shot, to
compare against the… See the full description on the dataset page: https://huggingface.co/datasets/bbidpa/flutter-full-examples-v1.Chinese-Braille-Dataset-Full-Tone
Chinese Braille Sentence Corpus (Full Tone)
📃 [Paper] •
💻 [Code] •
📖 [Passage corpus] •
🎬 [Demo]
The sentence-level half of the Braille–Chinese parallel corpus used in
"Vision-Braille: A Curriculum Learning Toolkit and Braille–Chinese Corpus for Braille
Translation" (EMNLP 2026 Main Conference).
Every Braille sequence here retains all tone markers (retention rate r = 100). This is
the source corpus: the tone-omission variants used for curriculum training are… See the full description on the dataset page: https://huggingface.co/datasets/Violet-yo/Chinese-Braille-Dataset-Full-Tone.FullStack-Bench
FullStack-Agent
Paper | Code | Dataset
Overview
This repository contains the FullStack-Bench dataset, introduced in the paper "FullStack-Agent: Enhancing Agentic Full-Stack Web Coding via Development-Oriented Testing and Repository Back-Translation".
In this paper, we propose FullStack-Agent, a unified system that combines a multi-agent full-stack development framework equipped with efficient coding and debugging tools (FullStack-Dev), an iterative self-improvement method… See the full description on the dataset page: https://huggingface.co/datasets/luzimu/FullStack-Bench.