The algorithm and the evaluation templates could benefit from a .dockerignore file. This helps speed up local testing as the build context becomes too large if there's a big .tar.gz (from an exported container) or a .git folder with neural network weights committed through LFS.
The .dockerignore file could comprise the following for starters:
The
algorithm
and theevaluation
templates could benefit from a.dockerignore
file. This helps speed up local testing as the build context becomes too large if there's a big.tar.gz
(from an exported container) or a.git
folder with neural network weights committed through LFS.The
.dockerignore
file could comprise the following for starters: