Opendoc1-Analysis-Recognition Dataset

Overview

The Opendoc1-Analysis-Recognition dataset is designed for tasks involving image-to-text, text classification, and image feature extraction. It contains images paired with class labels, making it suitable for vision-language tasks.

Dataset Details

• Modalities: Image
• Languages: English
• Size: Approximately 1,000 samples (n=1K)
• Tags: image, analysis, vision-language
• License: Apache 2.0

Tasks

This dataset can be used for the following tasks:

• Image-to-Text: Convert images into textual descriptions.
• Text Classification: Classify text associated with images.
• Image Feature Extraction: Extract features from images for further analysis.

Dataset Structure

The dataset is split into a training set with 232 rows. Each row contains:

• Image: The image file.
• Label: The class label associated with the image.

Usage

To use this dataset, you can download it from the Hugging Face Datasets Hub. The dataset files are approximately 443 MB in size, and the auto-converted Parquet files are around 464 MB.

Download Information

• Size of downloaded dataset files: 443 MB
• Size of the auto-converted Parquet files: 464 MB
• Number of rows: 386