Datasets:
dataset_info:
features:
- name: videoid
dtype: int64
- name: contentUrl
dtype: string
- name: duration
dtype: string
- name: page_dir
dtype: string
- name: name
dtype: string
splits:
- name: train
num_bytes: 2993227031
num_examples: 10727607
- name: validation
num_bytes: 1394310
num_examples: 5000
download_size: 1564635003
dataset_size: 2994621341
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
task_categories:
- text-to-video
- video-classification
size_categories:
- 10M<n<100M
Dataset Name: Video-10M
Description:
Video-10M is a dataset consisting of 10 million videos, each accompanied by metadata including video ID, content URL, duration, page directory, and name. The dataset covers a wide range of video content, from nature scenes and outdoor activities to culinary arts and urban landscapes. It provides a diverse collection of videos suitable for various video classification tasks.
Split Information:
- Train Split: Contains 10,727,607 samples.
- Validation Split: Contains 5,000 samples.
Features:
- videoid: Unique identifier for the video.
- contentUrl: URL or path to access the video content.
- duration: Duration of the video in hours, minutes, and seconds.
- page_dir: Directory or category associated with the video.
- name: Descriptive name or title for the video.
Usage:
The dataset can be used for a wide range of video classification tasks, including but not limited to content categorization, scene recognition, activity recognition, and object detection. Researchers, developers, and practitioners can leverage this dataset to train and evaluate video classification models across different domains and applications.
Download
from datasets import load_dataset, Dataset
#Load the dataset
dataset = load_dataset("Mouwiya/Video-10M")
Contact
Mouwiya S. A. Al-Qaisieh [email protected]