Dataset · starmpcc
Synthetic clinical notes with questions
starmpcc/Asclepius-Synthetic-Clinical-Notes
157k synthetic discharge summaries with question-answer pairs for training clinical LLMs.
What's inside
The dataset is structured in a clinical note-question-answer format for training clinical LLMs. Synthetic notes were created from PMC-Patients reports using GPT-3.5, then instruction-response pairs were generated for 157k synthetic discharge summaries. It covers 8 tasks, including named entity recognition; the language is English.
What you can do with it
- Fine-tune a clinical LLM on SFT
- Train clinical note summarization
- Evaluate QA on clinical texts
- Tasks
- Question answering, Summarization, Text generation
- Languages
- English
- Size
- 100K–1M rows
- Downloads this month
- 2.1K
- Likes
- 121
Models trained on it
- microsoft/MediPhi-InstructText generation 3K
- microsoft/MediPhiText generation 2.6K
- mradermacher/Sanad-1.0-i1-GGUF 2K
- TheBloke/Asclepius-13B-GGUFText generation 1.7K
- microsoft/MediPhi-ClinicalText generation 1K
- mradermacher/MediPhi-i1-GGUF 872
- mradermacher/Asclepius-Llama3-8B-i1-GGUF 835
- starmpcc/Asclepius-Llama2-7BText generation 670
How to load it
from datasets import load_dataset
ds = load_dataset("starmpcc/Asclepius-Synthetic-Clinical-Notes")