Dataset Viewer
The dataset viewer is not available for this split.
Rows from parquet row groups are too big to be read: 443.87 MiB (max=286.10 MiB)
Error code: TooBigContentError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Bridge Defect VLM Dataset (Trainable)
This dataset contains 150 bridge defect images with structured text annotations merged into a single description field.
Each sample includes:
- An image (
image) - A title (
title) – static string: "橋梁劣化檢測" - A full description prompt (
description) - A unique identifier (
uniq-id)
The dataset is designed to support vision-language model fine-tuning such as LLaVA, MiniGPT-4, LLaMA-Adapter, etc.
- Downloads last month
- 43