Dataset · FreedomIntelligence
Medical SFT reasoning from DeepSeek-R1
FreedomIntelligence/Medical-R1-Distill-Data
Medical tasks with R1 reasoning chains for model fine-tuning.
What's inside
An SFT dataset distilled from DeepSeek-R1 (Full Power) on verifiable medical tasks from HuatuoGPT-o1. Contains text question-answer pairs with reasoning; languages are English and Chinese, size ranges from 10k to 100k examples. Requests to DeepSeek-R1 were made via the native API; a separate Chinese version exists.
What you can do with it
- Fine-tuning medical LLMs to reason
- Initializing a model with R1 chains
- Evaluating QA on medical tasks
- Tasks
- Question answering, Text generation
- Languages
- English, Chinese
- Size
- 10K–100K rows
- Downloads this month
- 2.7K
- Likes
- 78
Models trained on it
- bartowski/allura-org_Q3-30B-A3B-Designant-GGUFText generation 2.4K
- mradermacher/MS3.2-24b-Angel-i1-GGUF 717
- mradermacher/Mistral-Small-24b-Sertraline-0304-i1-GGUF 711
- mradermacher/Q3-30B-A3B-Designant-i1-GGUF 675
- mradermacher/Q3-8B-Kintsugi-i1-GGUF 608
- mradermacher/Q3-30B-A3B-Designant-GGUF 603
- mradermacher/MS3.2-24b-Angel-GGUF 530
- bartowski/allura-org_Mistral-Small-24b-Sertraline-0304-GGUFText generation 512
How to load it
from datasets import load_dataset
ds = load_dataset("FreedomIntelligence/Medical-R1-Distill-Data")