117 lines
2.8 KiB
Markdown
117 lines
2.8 KiB
Markdown
# MRI-MCQA
|
|
|
|
## Overview
|
|
|
|
MRI-MCQA is a specialized benchmark composed of multiple-choice questions related to Magnetic Resonance Imaging (MRI). It evaluates AI models' understanding of MRI physics, protocols, image acquisition, and clinical applications.
|
|
|
|
## Task Description
|
|
|
|
- **Task Type**: Medical Imaging Knowledge Multiple-Choice QA
|
|
- **Input**: MRI-related question with multiple answer choices
|
|
- **Output**: Correct answer letter
|
|
- **Domain**: Medical imaging, MRI physics, radiology
|
|
|
|
## Key Features
|
|
|
|
- Specialized focus on MRI technology and applications
|
|
- Tests understanding of MRI physics and protocols
|
|
- Covers clinical MRI applications and sequences
|
|
- Designed for evaluating medical imaging AI systems
|
|
- Multiple-choice format for standardized evaluation
|
|
|
|
## Evaluation Notes
|
|
|
|
- Default configuration uses **0-shot** evaluation
|
|
- Evaluates on test split
|
|
- Simple accuracy metric
|
|
- No training split available
|
|
|
|
## Properties
|
|
|
|
| Property | Value |
|
|
|----------|-------|
|
|
| **Benchmark Name** | `mri_mcqa` |
|
|
| **Dataset ID** | [extraordinarylab/mri-mcqa](https://modelscope.cn/datasets/extraordinarylab/mri-mcqa/summary) |
|
|
| **Paper** | N/A |
|
|
| **Tags** | `Knowledge`, `MCQ`, `Medical` |
|
|
| **Metrics** | `acc` |
|
|
| **Default Shots** | 0-shot |
|
|
| **Evaluation Split** | `test` |
|
|
|
|
|
|
## Data Statistics
|
|
|
|
| Metric | Value |
|
|
|--------|-------|
|
|
| Total Samples | 563 |
|
|
| Prompt Length (Mean) | 457.23 chars |
|
|
| Prompt Length (Min/Max) | 259 / 888 chars |
|
|
|
|
## Sample Example
|
|
|
|
**Subset**: `default`
|
|
|
|
```json
|
|
{
|
|
"input": [
|
|
{
|
|
"id": "84f179d7",
|
|
"content": "Answer the following multiple choice question. The entire content of your response should be of the following format: 'ANSWER: [LETTER]' (without quotes) where [LETTER] is one of A,B,C,D.\n\nWhich cardiac chambers are typically imaged on the short-axis view?\n\nA) RA and RV\nB) RA and LA\nC) LA and LV\nD) RV and LV"
|
|
}
|
|
],
|
|
"choices": [
|
|
"RA and RV",
|
|
"RA and LA",
|
|
"LA and LV",
|
|
"RV and LV"
|
|
],
|
|
"target": "D",
|
|
"id": 0,
|
|
"group_id": 0,
|
|
"metadata": {}
|
|
}
|
|
```
|
|
|
|
## Prompt Template
|
|
|
|
**Prompt Template:**
|
|
```text
|
|
Answer the following multiple choice question. The entire content of your response should be of the following format: 'ANSWER: [LETTER]' (without quotes) where [LETTER] is one of {letters}.
|
|
|
|
{question}
|
|
|
|
{choices}
|
|
```
|
|
|
|
## Usage
|
|
|
|
### Using CLI
|
|
|
|
```bash
|
|
evalscope eval \
|
|
--model YOUR_MODEL \
|
|
--api-url OPENAI_API_COMPAT_URL \
|
|
--api-key EMPTY_TOKEN \
|
|
--datasets mri_mcqa \
|
|
--limit 10 # Remove this line for formal evaluation
|
|
```
|
|
|
|
### Using Python
|
|
|
|
```python
|
|
from evalscope import run_task
|
|
from evalscope.config import TaskConfig
|
|
|
|
task_cfg = TaskConfig(
|
|
model='YOUR_MODEL',
|
|
api_url='OPENAI_API_COMPAT_URL',
|
|
api_key='EMPTY_TOKEN',
|
|
datasets=['mri_mcqa'],
|
|
limit=10, # Remove this line for formal evaluation
|
|
)
|
|
|
|
run_task(task_cfg=task_cfg)
|
|
```
|
|
|
|
|