Scottie201/multimedia_dataset

Multimedia Dataset

This dataset contains both video and image files organized for machine learning tasks.

Dataset Structure

Scottie201/multimedia_dataset/
├── train/
│   ├── video_file_1.mp4
│   ├── video_file_2.MP4
│   ├── image_file_1.jpg
│   ├── image_file_2.PNG
│   ├── ...
│   └── metadata.csv
└── README.md

Supported Media Formats

Video Files (case-insensitive)

• mp4, avi, mov, webm, mkv, flv, wmv, m4v, 3gp, ogv

Examples: .mp4, .MP4, .avi, .AVI, .mov, .MOV, etc.

Image Files (case-insensitive)

• jpg, jpeg, png, gif, bmp, tiff, tif, webp, svg, ico

Examples: .jpg, .JPG, .png, .PNG, .gif, .GIF, etc.

Dataset Statistics

• Total files: 9 (2 videos, 7 images)
• Split: train
• File formats: Mixed multimedia formats
• Use cases: Computer vision, multimodal learning, media classification

Loading the Dataset

from datasets import load_dataset

# Load the full dataset
dataset = load_dataset("Scottie201/multimedia_dataset", data_dir="train")

# Filter by media type
video_dataset = dataset.filter(lambda x: x['media_type'] == 'video')
image_dataset = dataset.filter(lambda x: x['media_type'] == 'image')

print(f"Total samples: {len(dataset['train'])}")
print(f"Videos: {len(video_dataset['train'])}")
print(f"Images: {len(image_dataset['train'])}")

Metadata Format

The metadata.csv file contains the following columns:

• file_name: Name of the media file (preserves original case)
• media_type: Type of media ("video" or "image")
• timestamp: File modification timestamp
• file_size: File size in bytes
• extension: File extension (normalized to lowercase)
• text: Description or caption for the media file

Usage Examples

Working with Videos Only

from datasets import load_dataset

dataset = load_dataset("Scottie201/multimedia_dataset", data_dir="train")
videos = dataset.filter(lambda x: x['media_type'] == 'video')

for video in videos['train']:
    print(f"Video: {video['file_name']}")
    # Process video file...

Working with Images Only

from datasets import load_dataset

dataset = load_dataset("Scottie201/multimedia_dataset", data_dir="train")
images = dataset.filter(lambda x: x['media_type'] == 'image')

for image in images['train']:
    print(f"Image: {image['file_name']}")
    # Process image file...

Mixed Media Processing

from datasets import load_dataset

dataset = load_dataset("Scottie201/multimedia_dataset", data_dir="train")

for item in dataset['train']:
    if item['media_type'] == 'video':
        print(f"Processing video: {item['file_name']}")
        # Handle video
    elif item['media_type'] == 'image':
        print(f"Processing image: {item['file_name']}")
        # Handle image

Features

• Case-insensitive file extension matching: Both uppercase and lowercase extensions work
• Automatic media type detection: Files are automatically classified as video or image
• Comprehensive metadata: Includes file size, timestamps, and media type information
• Easy filtering: Use the media_type column to work with specific media types
• HuggingFace compatibility: Fully compatible with the datasets library

License

This dataset is released under the Apache 2.0 License.

Citation

If you use this dataset in your research, please cite:

@dataset{multimedia_dataset_2025,
  title={Multimedia Dataset},
  author={Scottie201},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/Scottie201/multimedia_dataset}
}

This dataset can be used with Hugging Face's datasets library and supports automatic media loading for both images and videos.