Giter Club home page Giter Club logo

kinetics-dataset's Introduction

Kinetics datasets

Kinetics is a collection of large-scale, high-quality datasets of URL links of up to 650,000 video clips that cover 400/600/700 human action classes, depending on the dataset version. The videos include human-object interactions such as playing instruments, as well as human-human interactions such as shaking hands and hugging. Each action class has at least 400/600/700 video clips. Each clip is human annotated with a single action class and lasts around 10 seconds.

The Kinetics project publications can be found here: https://deepmind.com/research/open-source/kinetics.

Download Videos

CVDF currently hosts the videos in the Kinetics-400 and Kinetics-700-2020 datasets.

Kinetics-400

The train/val/test splits are subdivided into many files. The lists of links to video files can be found here:

https://s3.amazonaws.com/kinetics/400/train/k400_train_path.txt

https://s3.amazonaws.com/kinetics/400/val/k400_val_path.txt

https://s3.amazonaws.com/kinetics/400/test/k400_test_path.txt

It is easy to obtain a specific split (e.g. train), by:

bash download.sh k400_train_path.txt

Then, extract *.tar.gz files by:

bash extract.sh k400_train_path.txt

The links/annotations can be found under the annotation subfolders:

https://s3.amazonaws.com/kinetics/400/annotations/train.csv

https://s3.amazonaws.com/kinetics/400/annotations/val.csv

https://s3.amazonaws.com/kinetics/400/annotations/test.csv

A readme file can be found in:

http://s3.amazonaws.com/kinetics/400/readme.md

Kinetics-600

The train/val/test splits are subdivided into many files. The lists of links to video files can be found here:

https://s3.amazonaws.com/kinetics/600/train/k600_train_path.txt

https://s3.amazonaws.com/kinetics/600/val/k600_val_path.txt

https://s3.amazonaws.com/kinetics/600/test/k600_test_path.txt

The links/annotations can be found under the annotation subfolders:

https://s3.amazonaws.com/kinetics/600/annotations/train.txt

https://s3.amazonaws.com/kinetics/600/annotations/val.txt

https://s3.amazonaws.com/kinetics/600/annotations/test.csv

A readme file can be found in:

http://s3.amazonaws.com/kinetics/600/readme.md

Kinetics-700-2020

The train/val/test splits are subdivided into many files. The lists of links to video files can be found here:

https://s3.amazonaws.com/kinetics/700_2020/train/k700_2020_train_path.txt

https://s3.amazonaws.com/kinetics/700_2020/val/k700_2020_val_path.txt

https://s3.amazonaws.com/kinetics/700_2020/test/k700_2020_test_path.txt

The links/annotations can be found under the annotation subfolders:

https://s3.amazonaws.com/kinetics/700_2020/annotations/train.csv

https://s3.amazonaws.com/kinetics/700_2020/annotations/val.csv

https://s3.amazonaws.com/kinetics/700_2020/annotations/test.csv

A readme file can be found in:

http://s3.amazonaws.com/kinetics/700_2020/K700_2020_readme.txt

Downstream annotations

We also host annotations for AVA-Kinetics and Countix, which both use Kinetics-700 videos.

To download annotations for AVA-Kinetics: https://s3.amazonaws.com/kinetics/700_2020/annotations/ava_kinetics_v1_0.tar.gz

To download annotations for countix: https://s3.amazonaws.com/kinetics/700_2020/annotations/countix.tar.gz

kinetics-dataset's People

Contributors

kinetics-cvdf avatar tylin avatar shoufachen avatar

Watchers

James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.