Keepsake Version control for machine learning

The Keepsake CLI sends anonymous analytics about commands you run.

You will be notified the first time you run keepsake.

Why we do this

It helps us:

1. Understand what commands people use and what's failing so we can prioritize work.
2. Understand what operating systems, Python, and Tensorflow/PyTorch versions people use so we can figure out what ones to support.
3. Figure out roughly how many people use Keepsake and what its growth rate is. We intend to build a business to support the open-source project, and this helps us get investment and so on. (This is the unspoken truth as to why most things gather analytics, but we want to be honest with you.)

What data is sent

• A random token for the machine (e.g. 9f3027bb-0eb8-917d-e5bf-c6c1bdb1fd0a)
• The subcommand you ran, without any options or arguments (e.g. keepsake run, not keepsake run python secretproject.py)
• The Keepsake version (e.g. 1.0.0)
• Your CPU architecture (e.g. amd64)
• Your operating system (e.g. linux)

Opting out

These analytics really help us, and we'd really appreciate it if you left it on. But, if you want to opt out, you can run this command:

keepsake analytics off

Or, you can use an environment variable:

export KEEPSAKE_NO_ANALYTICS=1