The goal of the Kinetics dataset is to help the computer vision and machine learning communities advance models for video understanding. Given this large human action classification dataset, it may be possible to learn powerful video representations that transfer to different video tasks.
The Kinetics-700-2020 dataset will be used for this challenge. Kinetics-700-2020 is a large-scale, high-quality dataset of YouTube video URLs which include a diverse range of human focused actions. The aim of the Kinetics dataset is to help the machine learning community create more advanced models for video understanding. It is an approximate super-set of both Kinetics-400, released in 2017, Kinetics-600, released in 2018 and Kinetics-700, released in 2019.
The dataset consists of approximately 650,000 video clips, and covers 700 human action classes with at least 700 video clips for each action class. Each clip lasts around 10 seconds and is labeled with a single class. All of the clips have been through multiple rounds of human annotation, and each is taken from a unique YouTube video. The actions cover a broad range of classes including human-object interactions such as playing instruments, as well as human-human interactions such as shaking hands and hugging.
More information about how to download the Kinetics dataset is available here.
YouTube serves as the mainstream alternative to traditional Indonesian television (Sinetron). High-production talk shows, reality vlogs, and independent journalism formats attract millions of daily viewers. The platform acts as a democratic space where independent creators compete directly with legacy media networks. TikTok: The Trend Accelerator
Digital 2026: Indonesia — DataReportal – Global Digital Insights
The massive viewership numbers have translated into a robust creator economy. Brands have shifted substantial advertising budgets from traditional television networks to digital video campaigns. Hyper-localized influencer marketing is now standard practice, with brands leveraging micro-influencers who speak local dialects (such as Javanese, Sundanese, or Balinese) to build authentic consumer trust. Bokep Cewek Hijab Gemoy Suka di Ewe Dari Belakang - INDO18
Known locally as a major cultural driver, TikTok is the birthplace of Indonesia’s viral music trends, dance challenges, and micro-comedy sketches. It acts as a launchpad for indie musicians and niche content creators who achieve overnight national fame.
Indonesian entertainment is a vibrant ecosystem blending traditional folklore with high-budget modern productions and a massive digital creator economy. The landscape is dominated by long-running TV dramas ( YouTube serves as the mainstream alternative to traditional
She wasn’t famous. Not yet. But as the sun set over Jakarta’s smoky skyline, and millions of other creators posted dance challenges, comedy skits, and prank videos, Dinda had found something rarer: a corner of the Indonesian internet that felt like home. And in a country built on a thousand islands, a thousand languages, and a million daily struggles, that was the most viral thing of all.
Emotional ballads (often called "Galau" songs) by artists like , , and TikTok: The Trend Accelerator Digital 2026: Indonesia —
The Indonesian digital landscape is experiencing an unprecedented boom, driven by affordable mobile internet, a massive youth population, and a culturally rich society. Indonesia boasts some of the world's highest engagement rates on platforms like YouTube, TikTok, and Instagram. The phrase "Indonesian entertainment and popular videos" captures a diverse ecosystem where traditional heritage meets hyper-modern internet culture.
In terms of film and popular video essays, is Indonesia's most successful export. Movies like Pengabdi Setan (Satan's Slaves) and KKN di Desa Penari became cultural phenomena.
Indonesian videos on YouTube have gained significant traction globally, with many videos going viral and racking up millions of views. Some of the most popular Indonesian videos on YouTube include:
This modern, electronic-heavy subgenre of traditional Dangdut music dominates TikTok and YouTube. Tracks by artists like Denny Caknan mix Javanese lyrics with infectious beats, driving viral dance challenges and soundtracking millions of user-generated short videos.
1. Possible to use ImageNet checkpoints?
We allow finetuning from public ImageNet checkpoints for the supervised track -- but a link to the specific checkpoint should be provided with each submission.
2. Possible to use optical flow?
Flow can be used as long as not trained on external datasets, except if they are synthetic.
3. Can we train on test data without labels (e.g. transductive)?
No.
4. Can we use semantic class label information?
Yes, for the supervised track.
5. Will there be special tracks for methods using fewer FLOPs / small models or just RGB vs RGB+Audio in the self-supervised track?
We will ask participants to provide the total number of model parameters and the modalities used and plan to create special mentions for those doing well in each setting, but not specific tracks.