Dataset · unsloth
Mini sample of radiological images with descriptions
unsloth/Radiology_mini
A compact multimodal slice (0.33%) from the ROCOv2-radiology medical dataset.
What's inside
The dataset contains between 1,000 and 10,000 pairs of medical images and text descriptions. It is a small sample, representing 0.33% of the full eltorio/ROCOv2-radiology corpus, prepared by the Unsloth team.
What you can do with it
- Rapid prototyping and debugging of multimodal medical pipelines
- Training test LoRA adapters for analyzing radiological scans
- Size
- 1K–10K rows
- Downloads this month
- 634
- Likes
- 50
Models trained on it
- rizkysulaeman/Qwen3-VL-8B-Vision-GRPO-HealthCareImage text → text 300
- YLX1965/llama3-2-11b-xray-v1Image text → text 258
- rizkysulaeman/Qwen3-VL-8B-Vision-HealthcareImage text → text 203
- 2kfi/medgemma-4B-it-fine-tuned-gguf 158
- gawadx1/katyusha-3.5-2BImage text → text 120
- rizkysulaeman/Qwen3-VL-8B-Vision-GRPO-HealthCare-Q8_0-GGUFImage text → text 68
- 0llheaven/Llama-3.2-11B-Vision-Radiology-miniImage text → text 21
- RantiRepo/Llama-3.2-11B-Vision-Radiology-testingImage text → text 12
How to load it
from datasets import load_dataset
ds = load_dataset("unsloth/Radiology_mini")