A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Exploring the Characteristics of Popular Deep Learning GitHub Repositories*
Deep learning (DL) repositories on GitHub platform have a significant growth following the successful application of DL for different domain tasks. How to capture widespread and continued attention from such many DL repositories becomes an urgent concern for researchers and practitioners of DL. While some studies have attempted to explore popular GitHub repositories, they mainly focused on traditional software or academic AI repositories on GitHub. DL repositories are different from traditional software or academic AI repositories, it is necessary to explore the characteristics of popular DL GitHub repositories. Therefore, through a case study of 652 popular and unpopular DL repositories, we study the growth patterns and characteristics of popular DL repositories and the impact of four types of factors (i.e., basic information, DL-related information, README information, and repositories’ owner information) on the popularity of DL repositories. We find that: i) different from the four growth patterns in the traditional software repositories, DL repositories have three growth patterns: Faster-then-Slower, Linear, and Explosive growth patterns; ii) various features exert diverse influences on the popularity of repositories within distinct domains; ii) basic information of the GitHub repositories (e.g., number of contributions, and number of topics) and README information (e.g., has-quickexample, has-photo) are more helpful in attracting attention compared with DL-related information.
Exploring the Characteristics of Popular Deep Learning GitHub Repositories*
Deep learning (DL) repositories on GitHub platform have a significant growth following the successful application of DL for different domain tasks. How to capture widespread and continued attention from such many DL repositories becomes an urgent concern for researchers and practitioners of DL. While some studies have attempted to explore popular GitHub repositories, they mainly focused on traditional software or academic AI repositories on GitHub. DL repositories are different from traditional software or academic AI repositories, it is necessary to explore the characteristics of popular DL GitHub repositories. Therefore, through a case study of 652 popular and unpopular DL repositories, we study the growth patterns and characteristics of popular DL repositories and the impact of four types of factors (i.e., basic information, DL-related information, README information, and repositories’ owner information) on the popularity of DL repositories. We find that: i) different from the four growth patterns in the traditional software repositories, DL repositories have three growth patterns: Faster-then-Slower, Linear, and Explosive growth patterns; ii) various features exert diverse influences on the popularity of repositories within distinct domains; ii) basic information of the GitHub repositories (e.g., number of contributions, and number of topics) and README information (e.g., has-quickexample, has-photo) are more helpful in attracting attention compared with DL-related information.
Exploring the Characteristics of Popular Deep Learning GitHub Repositories*
Zhou, Yiren (author) / Qiao, Yu (author) / Xu, Tao (author)
2023-11-17
543676 byte
Conference paper
Electronic Resource
English
| TIBKAT | 2015
Github Actions vs AWS Code Build for CI
| TIBKAT | 2021
Unifying learning object repositories in MACE
| DataCite | 2007
Deep Geologic Repositories - by N. T. Rempe (Editor)
| Online Contents | 2011