datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
text-dataset-tiny-code-script-py-format
USED of tahamajs/medicine_ds_persian for .parquet file
USED of Alijafarixcs2/persian-it-llama2-2k for .parquet file
USED of Abirate/english_quotes for .jsonl file
NEW FILES (05/12/2025)
NEW FILES (12/26/2025)
NEW FILES (02/15/2026)
Argimi-Ardian-Finance-10k-text-image
The ArGiMI Ardian datasets : text and images
The ArGiMi project is committed to open-source principles and data sharing.
Thanks to our generous partners, we are releasing several valuable datasets to the public.
Dataset description
This dataset comprises 34,000 financial annual reports, written in English, meticulously
extracted from their original PDF format to provide a valuable resource for researchers and developers in financial
analysis and natural language… See the full description on the dataset page: https://huggingface.co/datasets/artefactory/Argimi-Ardian-Finance-10k-text-image.Text2WetLab
Text2WetLab
Natural language → Opentrons OT-2 robot protocol benchmark
Text2WetLab evaluates whether large language model agents can translate plain-English wet-lab procedure descriptions into correct, safe, executable liquid-handling protocols for the Opentrons OT-2.
Reference episode — MuJoCo 3D physics render
Results (pass@1, 2026-10-04)
3 Claude models evaluated as Claude Code agents in Harbor sandboxes on Modal. All 21 trials passed the… See the full description on the dataset page: https://huggingface.co/datasets/EvanOLeary/Text2WetLab.TextVQA_GT_bbox
TextVQA validation set with grounding truth bounding box
The dataset used in the paper MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMs for studying MLLMs' attention patterns.
The dataset is sourced from TextVQA and annotated manually with ground-truth bounding boxes.
We consider questions with a single area of interest in the image so that 4370 out of 5000 samples are kept.
Citation
If you find our paper and code useful… See the full description on the dataset page: https://huggingface.co/datasets/jrzhang/TextVQA_GT_bbox.Text2Receipt
Text2Receipt
Messy free-text Hebrew income notes -> valid, complete Israeli fiscal documents (receipts & tax invoices).
Live demo (Space): yonilev/Text2Receipt
Dataset: yonilev/Text2Receipt
Dataset Creation
A synthetic corpus from a deterministic, rule-based generator plus a bounded LLM-paraphrase layer, so the ground truth is exact by construction.
Pipeline
Scenario sampling - category, issuer status, document type, client type, year, payment… See the full description on the dataset page: https://huggingface.co/datasets/yonilev/Text2Receipt.Text2CGL
Text2CGL: A Clean Zero-Leakage Dataset for Vietnamese Geometry to CGL
This dataset contains formal mappings from natural Vietnamese geometry problem texts to Constructive Geometry Language (CGL) S-expressions for automated differentiable diagram synthesis.
📌 Dataset Releases & Versions
v1.0 (Paper Benchmark): 3,797 instruction–CGL pairs (3,037 train, 379 validation, 381 test) across 1,045 root problem lineages, as reported in the DASA 2026 paper.
v1.1 (Current… See the full description on the dataset page: https://huggingface.co/datasets/minn4/Text2CGL.Chinese-Image-Text-Corpus-dataset
REILX/Chinese-Image-Text-Corpus-dataset
[ English | 中文 ]
Introduction
The REILX/Chinese-Image-Text-Corpus-dataset is a multimodal dataset that pairs Chinese textual data with corresponding images. This dataset is derived from the Chinese-Xinhua Dictionary Database, which includes idioms, single characters, words, and aphorisms.
Dataset Structure
The dataset is organized into the following categories:
Idioms: Traditional Chinese idioms with explanations and… See the full description on the dataset page: https://huggingface.co/datasets/REILX/Chinese-Image-Text-Corpus-dataset.Curated-Fox-News-Headlines-and-Full-Text
Curated Fox News Headlines and Full Text
This dataset contains a clean, curated collection of Fox News articles, including both headlines and full article text. It is designed for use in natural language processing (NLP) tasks such as sentiment analysis, summarization, topic classification, and media analysis.
📁 Dataset Format
Format: CSV
Encoding: UTF-8
Fields:
headline: The article title or headline
publish_date: Date the article was published (YYYY-MM-DD)
content:… See the full description on the dataset page: https://huggingface.co/datasets/crawlfeeds/Curated-Fox-News-Headlines-and-Full-Text.eis-text250
EIS-Text250: 1970s U.S. Environmental Impact Statements (text-only)
Per-page OCR/extraction text for 250 scanned 1970s U.S. federal
Environmental Impact Statements (EIS) from the Northwestern University
Library collection — the text-only companion to
Windsao/eis-subset50
(which carries full page images for a 50-doc subset). Built to test how current
models handle long, dense, historical government text: mean ~300 pages/doc,
1970s typewriter prose, OCR noise from degraded… See the full description on the dataset page: https://huggingface.co/datasets/Windsao/eis-text250.PDD3_text_rendered_v2