Visualize Histogram metrics in Wandb
- Dominant language
- Jupyter Notebook
- Stars
- 358
- Forks
- 43
- Avg merge
- 9h 2m
- Merged PRs (30d)
- 2
Description
pfl supports reporting metrics to Weights & biases with https://apple.github.io/pfl-research/reference/callback.html#pfl.callback.WandbCallback.
Currently we only report the line metrics printed out in console.
Wandb has support for visualizing histograms.
It can be very useful if we can make use of this feature to visualize histogram of a metric over the cohort of users.
This depends on having https://github.com/apple/pfl-research/issues/91 implemented such that we can use that data for wandb visualization.
Contributor guide
Research direction
Start with the documented pfl.callback.WandbCallback entry point and review issue #91, which this work depends on. Determine how the cohort metric data becomes available, then verify that the resulting metric histogram is reported for visualization in Weights & Biases.
Written by the indexing model from the issue text.
Assessment
- Domain
- data-visualization, machine-learning
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100