Browse · 266 datasets
Health Datasets
Openly licensed datasets for training, fine-tuning, and evaluating health AI — spanning clinical records, medical QA, NER, radiology, drug safety, genomics, and more.
HealthCareMagic 100k
100k conversations
100k real patient-doctor conversations from HealthCareMagic platform. Covers a wide range of medical conditions and symptoms.
Use case: Train medical chatbots and virtual assistants. Best real-world dataset for patient-doctor conversation modelling.
Medical Guidelines (Meditron Training Data)
46k guidelines
Curated medical clinical practice guidelines from major health organisations. Used to train Meditron.
Use case: Fine-tuning LLMs on authoritative medical guidelines. Grounding health AI in evidence-based clinical recommendations.
Medical-O1 Reasoning SFT
90k samples
Medical reasoning chain dataset with step-by-step clinical reasoning traces. 90k samples with detailed thought processes for medical questions.
Use case: Fine-tuning LLMs for chain-of-thought medical reasoning. Teach models to reason through clinical problems step by step.
MedMCQA
194k questions
194k multiple-choice questions from Indian medical entrance exams (AIIMS and NEET PG) across 21 subjects.
Use case: Large-scale medical knowledge evaluation. Best dataset for training models on broad medical facts.
MedQA (USMLE)
12.7k questions
Multiple-choice questions from US medical licensing exams (USMLE Steps 1, 2, 3).
Use case: The gold standard for evaluating clinical reasoning. Train models to pass medical board exams.
MIMIC-CXR
227k chest X-rays
Chest X-ray images paired with radiology reports from BIDMC. Largest public chest X-ray dataset with text.
Use case: Radiology report generation, chest X-ray finding classification, multimodal health models.
MIMIC-IV
40k+ patients
De-identified EHR data from Beth Israel Deaconess Medical Center ICU. Notes, labs, vitals, medications.
Use case: Pre-training clinical LLMs, mortality prediction, ICU outcome modelling, clinical note generation.
NCBI Disease
793 abstracts
PubMed abstracts annotated with disease mentions, normalised to MeSH or OMIM concepts.
Use case: Train disease NER models. Evaluate entity extraction pipelines for clinical text.
PubHealth (HealthFact)
11.8k claims
Public health claims labelled as true, false, mixture, or unproven. Sourced from fact-checking organisations.
Use case: Build health misinformation detectors. Evaluate claim verification for RAG systems.
PubMedQA
273k QA pairs
Yes/No/Maybe questions grounded in PubMed research abstracts. Tests evidence-based reasoning.
Use case: Train and test evidence-based medicine. Evaluate whether a model can read an abstract and answer a clinical question.
350k dataset health ar en th
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
350k dataset health ar en th tokenized
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
ABS DATA test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
ABS DATA train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
adaption medical qa diagnosis treatment aug
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
africa synth maternal health hepatitis b mtct dataset all
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
africa synth maternal health preeclampsia dataset all
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
africa synth mental health anxiety dataset all
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
africa synth mental health depression dataset all
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
africa synth mental health psychosis dataset seychelles
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
alexteboul diabetes health indicators dataset
varies
A ner dataset for health AI research and evaluation.
Use case: Fine-tune models for biomedical named entity recognition.
ancestry dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Ara Egy Medical QA
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
arabic medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
arabic medical qa MERGED MAQA MMMLU MI
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
ASR mental health luganda dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
bangla health related paraphrased dataset
varies
A ner dataset for health AI research and evaluation.
Use case: Fine-tune models for biomedical named entity recognition.
bangla medical qa cot and non cot
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
BC5CDR
1.5k PubMed articles
Chemical and disease mentions in PubMed articles. Annotated for NER and chemical-disease relation extraction.
Use case: Drug-disease relation extraction, pharmacovigilance pipelines, chemical NER.
BC5CDR chemical dev
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
BC5CDR chemical test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
BC5CDR chemical train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
BioASQ
5k+ QA pairs
Expert-curated biomedical QA from PubMed. Includes yes/no, factoid, list, and summary question types.
Use case: Biomedical IR systems, evidence synthesis, multi-type medical QA training.
biobert ner medical text
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
biomedical dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
biomedical instruction dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
biomedical novelty claim corpus
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
BioNLI test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
BioRED
600 abstracts
Biomedical relation extraction dataset with multi-type entity pairs: gene, disease, chemical, variant.
Use case: Multi-type biomedical relation extraction, knowledge graph construction, drug-gene-disease links.
chatgpt 4o medical qa dataset 50000
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
chemprot dev
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
chemprot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
chemprot train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
CheXpert
224k X-rays
224k chest X-rays from Stanford Medicine with 14 radiological findings labels, including uncertainty labels.
Use case: Chest X-ray classification, abnormality detection, radiology AI training.
chinese biomedical NER dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
chinese medical ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Chinese medical QA
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
clair health medical datasets
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Clinical Information Extraction
172 documents
Clinical information extraction dataset from MIT. Annotated clinical notes for structured information extraction.
Use case: EHR information extraction, clinical NER from free text, structured data extraction pipelines.
Clinical NLP benchmark dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
clinical nlp dataset raw
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
clinical nlp extraction
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
clinical nlp patient notes
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
clinical nlp test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
Clinical Trials History
545M rows
Comprehensive historical dataset of clinical trials from ClinicalTrials.gov with structured metadata and outcomes.
Use case: Clinical trial outcome prediction, protocol generation, trial eligibility matching, evidence synthesis.
ClinVar
1M+ variants
NCBI archive of relationships between genomic variation and human health with clinical significance labels.
Use case: Variant pathogenicity prediction, genomic report generation, precision medicine NLP.
CochranePLS 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
CochranePLS test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
CochranePLS train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
CORD-19
1M+ papers
1M+ scientific papers on COVID-19 and related coronaviruses with full-text and metadata.
Use case: Pre-training on infectious disease literature, evidence retrieval for RAG pipelines.
CT Scan Radiology Reports With Findings Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
CT Scan Radiology Reports Without Findings Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
custom dataset radiology
varies
A radiology dataset for health AI research and evaluation.
Use case: Train and evaluate models on radiology reports and imaging data.
dataset health
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
datashop medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
DDI Extraction
792 documents
Drug-drug interaction annotations from DrugBank and MedLine. Standard pharmacological NLP benchmark.
Use case: Drug safety NLP, pharmacovigilance, relation extraction for drug interaction detection.
DDI2013 dev
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
DDI2013 test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
DDI2013 train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
details Kabster BioMistral MedicalQA FT
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
drhast persian medical QA
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
dyspnea clinical notes
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
edition 0218 lavita medical qa shared task v1 toy readymade
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
eICU Collaborative Research Database
200k+ ICU stays
Multi-site ICU database covering 200k patient stays from 208 US hospitals. Richer demographic diversity than MIMIC.
Use case: ICU outcome prediction, treatment recommendation, multi-site clinical model training.
ELRC Medical V2
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
en ru health only dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
English Nuer Dinka Health Translation Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
gaming and mental health dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
HEAD-QA
6.8k questions
Spanish healthcare exam questions translated to English. Covers medicine, pharmacy, nursing, psychology, chemistry, biology.
Use case: Evaluate models across multiple healthcare disciplines. Good for multi-specialty health AI assessment.
health data factory datasets
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Health Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
health log extraction dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
health misinformation reddit dataset
varies
A ner dataset for health AI research and evaluation.
Use case: Fine-tune models for biomedical named entity recognition.
Health QA Finetune Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
health topic dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
HoC 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
HoC test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
HoC train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
huatuo medical qa sharegpt
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
ICD-10 Clinical Terminology
72.8k codes
ICD-10-CM and ICD-10-PCS diagnosis and procedure codes with descriptions. The standard for medical coding.
Use case: Automated ICD coding, clinical coding assistance, diagnosis classification from clinical notes.
indonesian medical ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Indonesian Regional Air Quality and Public Health Monitoring Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Instruction public health dataset
varies
A instruction tuning dataset for health AI research and evaluation.
Use case: Fine-tune general LLMs for health-specific instruction following.
iptacopan pubmed dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
JNLPBA
22k sentences
Biomedical NER covering DNA, RNA, cell types, cell lines, and proteins from MEDLINE abstracts.
Use case: Genomics NLP, biomedical entity recognition for biology-focused models.
khmer medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
leo pubmed dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
LitCovid 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
LitCovid test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
LitCovid train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
LOINC Clinical Terminology
99k codes
Logical Observation Identifiers Names and Codes — the universal standard for lab tests, vital signs, and clinical measurements.
Use case: EHR standardisation, lab result NLP, interoperability tasks, clinical code prediction.
MedDialog
1.1M conversations
1.1M medical dialogue datasets in English and Chinese between patients and doctors.
Use case: Train patient-facing medical chatbots, symptom checker dialogues, conversational health assistants.
medical arabic qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical cardiology qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical entity qa data
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Medical Health QA Articles Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical insurance qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Medical Meadow
160k samples
Diverse medical instruction-tuning dataset combining flashcards, wikidoc, stack exchange, and more.
Use case: Fine-tuning general LLMs for health tasks. Comprehensive instruction-tuning corpus for medical alignment.
Medical Meadow Flashcards
33k samples
Medical flashcard Q&A pairs covering pharmacology, anatomy, physiology, and clinical medicine.
Use case: Medical knowledge instruction tuning. Good for teaching factual medical knowledge in a Q&A format.
medical multiple choice qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Medical NER
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MEDICAL NER
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner bleurt separated
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner instruction data
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner instruction data v2
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner instruction data v3
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner sft
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical ner vie
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical prescription ner 2026
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical privacy ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical QA
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa dataset merged
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa fr v0.1
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa instruction zhtw
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Medical QA Mistral7B Finetuning
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa ru data
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa ru data
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa sft ru
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa shared task v1 all
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa shared task v1 toy
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medical qa tokenized
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Medical Questions Pairs
3k pairs
Pairs of similar medical questions with labels indicating whether they are duplicates. Sourced from medical forums.
Use case: Medical question deduplication, semantic similarity for health chatbots, QA retrieval systems.
Medical Reasoning SFT (GPT-OSS 120B)
200k samples
200k medical reasoning SFT samples generated by a 120B open-source model. High-quality chain-of-thought annotations.
Use case: Large-scale medical instruction tuning. Best for training models on high-quality clinical reasoning traces.
Medical Transcription Dataset
5k transcriptions
Medical transcription samples across 40 specialties. Real clinical documentation covering various medical specialties.
Use case: Medical transcription NLP, clinical specialty classification, medical note generation across specialties.
medicalner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medicalNER
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medicalNer CAS annotations combined
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medicalNer CAS annotations combined v2
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medicalNer CAS annotations combined v2 new split
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedicalNER fr
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedicalNER Fr
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedicalNER Fr e3c guidelines
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedicalQA
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedicalQA part7 en
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedicalQA TR
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedicalQAChunkRetrieval
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
meditron clinical guidelines
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
medmcqa bonvoyage 39995 recall 0.5 qwen3 4b medical qa ht
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
medMCQA test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
MedMentions
4.4k abstracts
PubMed abstracts with mentions linked to UMLS concepts. The largest public biomedical NER dataset.
Use case: Fine-grained biomedical NER, medical knowledge graph population, concept normalisation.
MedNLI
14.1k pairs
Natural language inference pairs derived from MIMIC-III clinical notes. Tests clinical text entailment.
Use case: Train models to understand contradiction and entailment in clinical documentation. Key for EHR QA systems.
medQA 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
medQA test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
MedQA train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
MedQuAD
47k QA pairs
47k medical QA pairs from 12 NIH websites covering diseases, drugs, treatments, and research.
Use case: Patient-facing medical chatbots, health information retrieval, consumer health QA.
medquad medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MedShare ClinicalNLP
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
MedShare ClinicalNLP Raw
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
Menstrual Health Awareness Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
mental health chat dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
mental health chatbot dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
mental health chatbot dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Mental Health Classification Dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
mental health conversational dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Mental Health Counseling Conversations
3.5k conversations
High-quality mental health counseling conversation pairs between patients and therapists. Carefully curated for safety.
Use case: Fine-tune LLMs for empathetic mental health support, crisis response, and therapeutic dialogue systems.
mental health dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Mental health dataset4Fine Tuning
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Mental Health Knowledge Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
mental health survey dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Mental Health Text Classification Dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
Mental Health Text Classification Dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
Mental Health Text Classification Dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
Mental Health Text Classification Dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
Mental Health Text Classification Dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
Mental Health Text Classification Dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
Migraine Acute Treatment Plan Correction and Self Correction Trajectory Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
mimic clinical notes 269990
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
mimic clinical notes from og
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
mimic clinical notes gte2 from mimic note preprocessed
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
mimic clinical notes gte3 from mimic note preprocessed
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
mimic clinical notes v2
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
mimic medical imaging qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
mmlu medical filtered
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Monolingual health dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MRI Radiology Reports With Findings Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MRI Radiology Reports Without Findings Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
MS²
470k papers
Multi-document summarization dataset from medical systematic reviews. Pairs multiple RCT abstracts with review summaries.
Use case: Evidence synthesis, systematic review automation, multi-document clinical summarization.
MS2 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
MS2 test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
MS2 train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
MTSample test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
multimodal mental health dataset MMH
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
My pubmed dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
myanmar medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
NCBI disease dev
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
NCBI disease test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
NCBI disease train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
NCD Instruct Tuning Medical QA Indonesian
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Nepali SCRAP Health Instruction Dataset
varies
A instruction tuning dataset for health AI research and evaluation.
Use case: Fine-tune general LLMs for health-specific instruction following.
NER medical reports
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
NER medical reports tokenized deid roberta i2b2
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
new dataset radiology
varies
A radiology dataset for health AI research and evaluation.
Use case: Train and evaluate models on radiology reports and imaging data.
new radiology dataset
varies
A radiology dataset for health AI research and evaluation.
Use case: Train and evaluate models on radiology reports and imaging data.
NIH Chest X-ray14
112k images
112k frontal chest X-rays with 14 disease labels from NIH Clinical Center. Classic radiology AI benchmark.
Use case: Chest pathology classification, pre-training visual encoders for multimodal health models.
paza health audio dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
pelacarsen pubmed dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
pet health symptoms dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
pet health symptoms dataset
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
PhysioNet ECG Databases
100k+ recordings
ECG signal databases including MIT-BIH arrhythmia, PTB-XL. Annotated for rhythm and morphology.
Use case: Arrhythmia detection, ECG classification, multimodal cardiac AI combining signals and clinical text.
PLOS 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
PLOS test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
PLOS train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
Pregnancy Medication Constraint Evaluation Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Primary Care Prescription Constraint Evaluation Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
PubMed Abstracts
30M+ abstracts
The complete PubMed archive of biomedical and life sciences journal abstracts. Updated daily.
Use case: Pre-training or continuing pre-training of biomedical LLMs. Building medical search and RAG indexes.
PubMed Cancer NLP Textual Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
PubMed Central Open Access
5M+ full-text papers
Full-text open access papers from PubMed Central covering all biomedical specialties.
Use case: Pre-training or domain adaptation of biomedical LLMs. Building knowledge-rich RAG indexes for health AI.
pubmed dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
PubMed MultiLabel Text Classification Dataset MeSH
varies
A nli dataset for health AI research and evaluation.
Use case: Evaluate natural language inference on biomedical text.
PubMed Summ 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
pubmed sycophancy dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
pubmed sycophancy dataset dpo
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
PubMed-RCT
4.5M sentences
Sentences from PubMed RCT abstracts labelled by role: background, objective, methods, results, conclusions.
Use case: Abstract structure prediction, evidence extraction from clinical trials, structured summary generation.
pubmeddataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
pubmedqa 1shot test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
pubmedqa test
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
PubmedQA train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
PubmedSumm train
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
radiology dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
radiology dataset
varies
A ner dataset for health AI research and evaluation.
Use case: Fine-tune models for biomedical named entity recognition.
radiology dataset 9
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
radiology detection dataset
varies
A radiology dataset for health AI research and evaluation.
Use case: Train and evaluate models on radiology reports and imaging data.
radiology recognition dataset
varies
A radiology dataset for health AI research and evaluation.
Use case: Train and evaluate models on radiology reports and imaging data.
radiology vlm dataset
varies
A radiology dataset for health AI research and evaluation.
Use case: Train and evaluate models on radiology reports and imaging data.
raft dataset pubmed
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
raft dataset pubmed 2
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
raw medical NER
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Real Scarp Nepali Health FAQ Complile Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
REAL SCRAP HEALTH COMPILE ROMANIZED DATASET
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
rntc clinical pubmed translate
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
rntc nuner pubmed french
varies
A ner dataset for health AI research and evaluation.
Use case: Fine-tune models for biomedical named entity recognition.
rntc pubmed fr clinical quality annotations 5
varies
A clinical records dataset for health AI research and evaluation.
Use case: Train models on real clinical documentation and EHR data.
rntc pubmed preprocess
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
russian medical qa synthetic
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
SIDER
1.4k drugs
Side effect and indication database covering 1427 drugs with 5880 adverse effects from drug package inserts.
Use case: Adverse drug event prediction, drug labelling NLP, pharmacovigilance model training.
STXBP1 PubMed Central Multimodal Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
swedish medical ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
swedish medical ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
swedish medical ner
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
syed med medical qa synthetic 4k
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
tokenized medical NER
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
vietnamese health dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
vietnamese medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
vietnamese medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
vietnamese medical qa
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
VQA-RAD
3.5k QA pairs
Clinician-generated QA pairs on radiology images (CT, MRI, chest X-ray). Tests visual medical reasoning.
Use case: Train and evaluate multimodal radiology AI. Benchmark for medical visual question answering.
Wiki Medical Terms
6.8k entries
Medical terms and their definitions scraped from Wikipedia. Clean, structured glossary-style entries.
Use case: Medical terminology fine-tuning, glossary-grounded RAG, medical concept explanation training.
X Ray Radiology Reports With Findings Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
X Ray Radiology Reports Without Findings Dataset
varies
A medical qa dataset for health AI research and evaluation.
Use case: Evaluate and train models on medical question answering tasks.
Xray OPG Dental Radiology Reports Without Findings Dataset
varies
A radiology dataset for health AI research and evaluation.
Use case: Train and evaluate models on radiology reports and imaging data.