Implementation:Open compass VLMEvalKit MedqbenchCaptionDataset

Field	Value
source	VLMEvalKit
domain	Vision, Benchmarking, Medical Imaging, Captioning

Overview

Benchmark dataset implementation for MedQ-Bench medical image captioning evaluation in VLMEvalKit.

Description

MedqbenchCaptionDataset inherits from ImageBaseDataset and implements the MedQ-Bench Caption benchmark for evaluating medical image description capabilities. The TYPE field is set to 'Caption'. It supports dev and test splits (MedqbenchCaption_dev, MedqbenchCaption_test) and includes a custom MedQBench_Caption_Scorer class for scoring generated medical captions.

Usage

Registered in vlmeval/dataset/__init__.py and invoked through build_dataset() by benchmark name.

Code Reference

Source: vlmeval/dataset/medqbench_caption.py, Lines: L1-457
Import: from vlmeval.dataset.medqbench_caption import MedqbenchCaptionDataset

Signature:

class MedqbenchCaptionDataset(ImageBaseDataset):
    TYPE = 'Caption'
    DATASET_URL = {...}
    DATASET_MD5 = {...}
    ...

I/O Contract

Direction	Description
Inputs	TSV dataset file with medical images and captioning tasks
Outputs	Evaluation results DataFrame with caption quality scores

Usage Examples

from vlmeval.dataset import build_dataset
dataset = build_dataset('MedqbenchCaption')

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment