Kaggle Scripts enable you to run R, Python, Julia, or R Markdown code directly on competition datasets.
The scripts run in Docker containers for simplicity and reproducibility. These Docker containers are automatically built from Kaggle’s R and Python Github repositories. If you’d like to add another library, make a pull request on the appropriate repository or ask us on the corresponding R and Python forum threads.
The script computations currently have the following restrictions:
|Max execution time||10 minutes|
|Max file output||500 MB|
|Memory limit||8 GB|
Input data files are stored in the read-only
Your scripts execute with
/kaggle/working set as the current working directory. It’s best to read the input files as relative paths (e.g. in R use
train <- read.csv("../input/train.csv").
Source: Google Alert for ML