Skip to content

Data Version Control

Jump to bottom
David N. Palacio edited this page May 17, 2022 · 4 revisions

Adding new files or folders. Please execute the commands in the following order:

  1. dvc add
  2. git add
  3. git commit
  4. git push
  5. dvc push

Updating a tracked file:

  1. dvc commit xxx.dvc
  2. dvc push
  3. git add xxx.dvc
  4. git commit -m
  5. git push

To check if the cache is synchronized: dvc status

DVC Policy

Once a dataset is created and pushed into DVC. If you are not actively using the dataset for experimentation, then make sure to erase it from /scratch folder (or /main folder in the container) so that we avoid space issues.

semeru-dvc

This is a dedicated repo for semeru benchmarks. This is the only repo for data version control. Here we include not only benchmarks but also statistical strategies to explore data and analyze results.

Clone this wiki locally