-
Notifications
You must be signed in to change notification settings - Fork 72
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Could experiment/checkpoint tables just be pandas dataframes? #499
Comments
Replicate looks great! I just happened across this issue and thought I would share some experience I had with metric logging. I would recommend using xarray instead of pandas dataframes; xarray was born out of the awkwardness that was extending pandas dataframes beyond two dimensions. This makes it much easier to do things like align or concatenate metrics across experiments, handle the difference between train/test and batch/epoch level-metrics. All of these tasks benefit from xarray's usage of coordinate systems. I made a live-logging/plotting library that makes use of xarray to this effect: https://github.com/rsokl/noggin
These are just for reference; I don't plan on maintaining noggin any further, so I don't mean to recommend that you actually leverage noggin (although there are probably nuggets to be taken from it). |
This is something neat that this does: https://github.com/richardliaw/track
The text was updated successfully, but these errors were encountered: