Searching runs
In this exercise, you will query experiment runs from multiple Unicorn experiments and return only runs that meet a certain desired criteria. This is helpful during the ML lifecycle if you need to compare runs data.
First you will create a filter string to capture runs for R-squared metric greater than .70
. Using the function from the mlflow
module that searches runs, you will then order them in descending order and search only the experiments "Unicorn Sklearn Experiments"
and "Unicorn Other Experiments"
.
The experiments have already been created in MLflow along with the R-squared metrics. The MLflow module will be imported.
This exercise is part of the course
Introduction to MLflow
Exercise instructions
- For variable
r_squared_filter
, create a filter string to capture"r2_score"
metrics"> .70"
. - Search runs for experiments
"Unicorn Sklearn Experiments"
and"Unicorn Other Experiments"
. - Order the results of R-squared in descending order.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Create a filter string for R-squared score
r_squared_filter = "____.____ > .70"
# Search runs
mlflow.____(experiment_names=["____", "____"],
filter_string=____,
order_by=["____.____ ____"])