Data Versioning Control
Created on 2023-05-12T20:19:17-05:00
File storage
Acts as a second layer of commands on top of (but not directly tied to) Git.
You use a separate checkout, push, and pull command to move files around to storage servers.
Files are read and written to external services (like S3-compatible ones) while pointers are left behind in the .dvc folder. Ostensibly you then have to commit the .dvc folder after such changes.
Make replacement
Includes a "pipeline" feature that is supposed to be "makefiles for machine learning."
Supposed to associate some input and output data; i didn't look super close in to the pipeline mechanism.