datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
klej-nkjp-nerVoxpopuli_NER
VoxPopuli_NER
VoxPopuli-NER is derived from the VoxPopuli corpus and specifically enhanced for
Named Entity Recognition (NER) tasks focusing on political and geographical entities.
It includes 879 audio samples, annotated with 2469 unique entity types. The dataset consists of the English part of the test set of VoxPopuli.
See full details in the WhisperNER paper.
citation
If you find this usful, please cite the following works:
@article{ayache2024whisperner… See the full description on the dataset page: https://huggingface.co/datasets/aiola/Voxpopuli_NER.cn_ner来源 https://github.com/liucongg/NLPDataSet
从网上收集数据，将CMeEE数据集、IMCS21_task1数据集、CCKS2017_task2数据集、CCKS2018_task1数据集、CCKS2019_task1数据集、CLUENER2020数据集、MSRA数据集、NLPCC2018_task4数据集、CCFBDCI数据集、MMC数据集、WanChuang数据集、PeopleDairy1998数据集、PeopleDairy2004数据集、GAIIC2022_task2数据集、WeiBo数据集、ECommerce数据集、FinanceSina数据集、BoSon数据集、Resume数据集、Bank数据集、FNED数据集和DLNER数据集等22个数据集进行整理清洗，构建一个较完善的中文NER数据集。
数据集清洗时，仅进行了简单地规则清洗，并将格式进行了统一化，标签为“BIO”。
处理后数据集详细信息，见数据集描述。
数据集由NJUST-TB一起整理。
由于部分数据包含嵌套实体的情况，所以转换成BIO标签时，长实体会覆盖短实体。
数据… See the full description on the dataset page: https://huggingface.co/datasets/ttxy/cn_ner.fire-financial-ner-extractionThis dataset can be used for benchmarking LLM Structured Outputs via the code here:
https://github.com/cleanlab/structured-output-benchmark/
vc-deal-flow-signal
Startup GitHub Engineering Velocity Panel
A longitudinal dataset of public GitHub engineering-activity signals for venture-backed startups. It is published under CC BY 4.0 for reproducible research, data journalism, and analysis of alternative data in venture capital.
219 startup-period observations
55 unique startups
18 sectors
4 quarterly periods: Q3 2025, Q4 2025, Q1 2026, and Q2 2026
No missing values in the primary table
Version: 1.0.0
The 219 rows are startup-period… See the full description on the dataset page: https://huggingface.co/datasets/the-data-nerd/vc-deal-flow-signal.ner_datasetbioleaflets-biomedical-ner
Dataset Card for BioLeaflets Dataset
Dataset Summary
BioLeaflets is a biomedical dataset for Data2Text generation. It is a corpus of 1,336 package leaflets of medicines authorised in Europe, which were obtained by scraping the European Medicines Agency (EMA) website.
Package leaflets are included in the packaging of medicinal products and contain information to help patients use the product safely and appropriately.
This dataset comprises the large majority (∼ 90%) of… See the full description on the dataset page: https://huggingface.co/datasets/ruslan/bioleaflets-biomedical-ner.resume_ner中文 resume ner 数据集， 来源: https://github.com/luopeixiang/named_entity_recognition 。
数据的格式如下，它的每一行由一个字及其对应的标注组成，标注集采用BIOES，句子之间用一个空行隔开。
美 B-LOC
国 E-LOC
的 O
华 B-PER
莱 I-PER
士 E-PER
我 O
跟 O
他 O
谈 O
笑 O
风 O
生 O
效果
不同模型的效果对比：
Bert-tiny 结果
model
precision
recall
f1-score
support
BERT-tiny
0.9490
0.9538
0.9447
全部
BERT-tiny
0.9278
0.9251
0.9313
使用 100 train
注：
后面再测试，BERT-tiny(softmax) + 100 训练样本，暂时没有复现 0.9313 的结果，最好结果 0.8612
BERT-tiny +… See the full description on the dataset page: https://huggingface.co/datasets/ttxy/resume_ner.Financial-NER-NLPDataset Card for Financial-NER-NLP
Dataset Summary
The Financial-NER-NLP Dataset is a derivative of the FiNER-139 dataset, which consists of 1.1 million sentences annotated with 139 XBRL tags. This new dataset transforms the original structured data into natural language prompts suitable for training language models. The dataset is designed to enhance models’ abilities in tasks such as named entity recognition (NER), summarization, and information extraction in the financial domain.
The… See the full description on the dataset page: https://huggingface.co/datasets/Josephgflowers/Financial-NER-NLP.fever-ner
FEVER Entity Retrieval Benchmark
Frozen benchmark for evaluating retrieval methods on the BEIR FEVER dataset (5.4M Wikipedia articles, 6,666 test queries). All data is pre-built so you can test a new method without re-running BM25 or dense retrieval.
Files
Core benchmark data (for testing new methods)
File
Size
What it is
beir_pool.json
31 MB
BM25 top-100 candidate pool (k1=1.2, b=0.75). 6,666 queries, each with 100 candidate docids +… See the full description on the dataset page: https://huggingface.co/datasets/Kim-el/fever-ner.globalise_NER_token_classification_dataset
Dataset Card for Dataset Name
The globalise_NER_token_classification dataset is a fine-grained dataset for the training of token-classification NER models on Dutch East-India Company texts (17th to 18th century).
Dataset Details
Dataset Description
The dataset provides 15 fine-grained labels detailing activities and people of the Dutch East-India Company (VOC), and can be used to train NER token-classification models for the
period 17th-18th century and the… See the full description on the dataset page: https://huggingface.co/datasets/globalise/globalise_NER_token_classification_dataset.PII-NERDataset Card for NER PII Extraction Dataset
Dataset Summary
This dataset is designed for training and evaluating Named Entity Recognition (NER) models focused on extracting Personally Identifiable Information (PII) from text. It includes a variety of entities such as names, addresses, phone numbers, email addresses, and identification numbers. The dataset is suitable for tasks that involve PII detection, compliance checks, and data anonymization.
Supported Tasks and Leaderboards
Named Entity… See the full description on the dataset page: https://huggingface.co/datasets/Josephgflowers/PII-NER.atis-ner-turkishThe ATIS (Airline Travel Information System) Dataset includes spoken queries (i.e., utterances) annotated for the task of slot filling in conversational systems.
This dataset, ATISNER, includes airline spoken queries translated from English to Turkish, customized for Named Entity Recognition.
Train and test splits include 4,978 and 890 sentences, respectively.
Translations are provided by the following study.
Şahinuç, F., Yücesoy, V., & Koç, A. (2020). Intent Classification and Slot Filling… See the full description on the dataset page: https://huggingface.co/datasets/ctoraman/atis-ner-turkish.wojood-arabic-ner
Wojood Arabic NER (Sample, Processed)
This dataset is a processed version of the sample from the Wojood Arabic NER dataset, developed by SinaLab.The original dataset is licensed under the MIT License, and this processed sample is shared under the same terms.
⚠️ Note: This is not the full Wojood corpus. The full dataset must be requested directly from the original authors (Prof. Mustafa Jarrar and team).
Dataset
Language: Arabic
Task: Named Entity Recognition (NER)… See the full description on the dataset page: https://huggingface.co/datasets/AhmedNabil1/wojood-arabic-ner.ner_movieMovie ner dataset
hard-nerve-dcb861
hard-nerve-dcb861
Synthetic sensors test data: 53 rows in data.csv.
All values are randomly generated fictional examples, not real observations, products, or user activity. Intended only for CSV loading and pipeline tests; not suitable for scientific or business conclusions. Columns are sampled independently and do not model real-world correlations.
Fields
sample_id: random identifier for this generated sample.
row_id: sequential row number starting at 1.… See the full description on the dataset page: https://huggingface.co/datasets/PaulSanchez/hard-nerve-dcb861.nervous-section-ccc5e3
nervous-section-ccc5e3
Synthetic products test data: 32 rows in data.csv.
All values are randomly generated fictional examples, not real observations, products, or user activity. Intended only for CSV loading and pipeline tests; not suitable for scientific or business conclusions. Columns are sampled independently and do not model real-world correlations.
Fields
sample_id: random identifier for this generated sample.
row_id: sequential row number starting at 1.… See the full description on the dataset page: https://huggingface.co/datasets/Indigo-Owen/nervous-section-ccc5e3.NER-benchmark-750
Crypto NER Benchmark
The crypto world has long awaited a robust NER benchmark and distinguished NER model, hindered by the unique challenges of the crypto realm. The space is characterized by sophisticated terminology, emotionally charged discourse, meme-driven content, and often misleading project names (e.g., NO, MOVE, DOGE). In response to this gap, the Cyber.co team has developed a comprehensive NER benchmark dataset, pioneering the first standardized evaluation framework in… See the full description on the dataset page: https://huggingface.co/datasets/cyberco/NER-benchmark-750.nervous-crew-48b38a
nervous-crew-48b38a
Synthetic weather test data: 46 rows in data.csv.
All values are randomly generated fictional examples, not real observations, products, or user activity. Intended only for CSV loading and pipeline tests; not suitable for scientific or business conclusions. Columns are sampled independently and do not model real-world correlations.
Fields
sample_id: random identifier for this generated sample.
row_id: sequential row number starting at 1.… See the full description on the dataset page: https://huggingface.co/datasets/Meridian-Sable/nervous-crew-48b38a.nervous-budget-86eac7
nervous-budget-86eac7
Synthetic products test data: 59 rows in data.csv.
All values are randomly generated fictional examples, not real observations, products, or user activity. Intended only for CSV loading and pipeline tests; not suitable for scientific or business conclusions. Columns are sampled independently and do not model real-world correlations.
Fields
sample_id: random identifier for this generated sample.
row_id: sequential row number starting at 1.… See the full description on the dataset page: https://huggingface.co/datasets/velvetThomas/nervous-budget-86eac7.Financial-NER-NLPDataset Card for Financial-NER-NLP
Dataset Summary
The Financial-NER-NLP Dataset is a derivative of the FiNER-139 dataset, which consists of 1.1 million sentences annotated with 139 XBRL tags. This new dataset transforms the original structured data into natural language prompts suitable for training language models. The dataset is designed to enhance models’ abilities in tasks such as named entity recognition (NER), summarization, and information extraction in the financial domain.
The… See the full description on the dataset page: https://huggingface.co/datasets/qyc777/Financial-NER-NLP.uzbek_NER
Uzbek NER Gold
Uzbek NER Gold is a token-level named entity recognition dataset for Uzbek. The dataset is distributed as a UTF-8 TSV file and uses BIO tagging for named entities.
Dataset Summary
Dataset ID: uznlp-uz/uzbek_NER
Language: Uzbek (uz)
Rows: 59,569 token rows
Columns: 5
Sentences: 4,176
Split: train
Format: UTF-8 TSV
Data file: Uzbek_NER_Gold.tsv
License: CC BY 4.0
Data Fields
Field
Description
Sentence
Sentence identifier.… See the full description on the dataset page: https://huggingface.co/datasets/uznlp-uz/uzbek_NER.bg_ner_bsnlp
Dataset Card for Bulgarian Named Entity Recognition. Initial dataset is taken from Balto-Slavic NLP shared task and is further transformed in the format appropriate for token classification. The instances are randomized and splitted into train and test splits.
Dataset Summary
This dataset is initially created for the BSNLP Shared Task 2019 and reported in the conference paper "The Second Cross-Lingual Challenge on Recognition, Normalization, Classification, and Linking of… See the full description on the dataset page: https://huggingface.co/datasets/usmiva/bg_ner_bsnlp.ner_dataBIS-Speeches-NER-dataset
BIS Speeches NER dataset
BIS Speeches NER dataset is a large-scale annotated dataset of central bank speeches (1996–2024) curated for Named Entity Recognition (NER).
It was developed to support the fine-tuning of Central Bank BERT and related models for extracting structured metadata from central bank communication.
Models Trained on This Dataset
Central Bank NER – fine-tuned NER model for AUTHOR, POSITION, AFFILIATION.
Dataset Schema
Each row… See the full description on the dataset page: https://huggingface.co/datasets/bilalzafar/BIS-Speeches-NER-dataset.synthetic-NER-dataset-v1
Tanaos NER Training Dataset
This dataset was created synthetically by Tanaos with the Artifex Python library.
The dataset is designed to train and evaluate Named Entity Recognition (NER) systems — models that identify and classify named entities in text into predefined categories such as PERSON, ORG, LOCATION, DATE, and more. It can be used to train NER models from scratch or fine-tune existing ones.
Our flagship NER model, tanaos-NER-v1, was trained on this dataset.… See the full description on the dataset page: https://huggingface.co/datasets/tanaos/synthetic-NER-dataset-v1.nervous-radio-8ab78b
nervous-radio-8ab78b
Synthetic products test data: 45 rows in data.csv.
All values are randomly generated fictional examples, not real observations, products, or user activity. Intended only for CSV loading and pipeline tests; not suitable for scientific or business conclusions. Columns are sampled independently and do not model real-world correlations.
Fields
sample_id: random identifier for this generated sample.
row_id: sequential row number starting at 1.… See the full description on the dataset page: https://huggingface.co/datasets/Velvet-Michael/nervous-radio-8ab78b.eval-gliner2-ner-fin-dice-soft-20260709ner_quechua_iic
Dataset Card for WikiANN
Dataset Summary
NER_Quechua_IIC is a named entity recognition dataset consisting of dictionary texts provided by the Peruvian Ministry of Education, annotated with LOC (location), PER (person) and ORG (organization) tags in the IOB2 format.
Supported Tasks and Leaderboards
named-entity-recognition: The dataset can be used to train a model for named entity recognition in Quechua languages.
ner-email
Overview:
This dataset is augmented through llama-3.1:8b. The pourpose is to finetune llm for token classification i.e Email in our case.
Following tags are present in dataset:
full_name : 1
email : 2
gender : 3
city : 4
country : 5