https://kedro.org/ logo
Join Slack
Powered by
Hey team, I would like to understand more about ml...
# questions
s
Hey team, I would like to understand more about mlflow setting in kedro. In 
mlflow.yml
, we have some parameters already. Like 
mlflow_tracking_uri
, 
experiment name
, ... But how can I define an 
artifact_uri
, which is in 
s3
? Can I simply add a parameter into 
mlflow.yml
, under server: 
artifact_uri: <s3://myuri>
? Another thing, how could I load parameters from other 
.yml
file into 
mlflow.yml
? Thank you in advance!
h
Someone will reply to you shortly. In the meantime, this might help:
r
I haven't worked directly with MLflow artifacts on S3 myself, but this discussion might help: https://github.com/Galileo-Galilei/kedro-mlflow/issues/15
y
If you refer to the artifact_uri specific to an experiment, you can specify it in :
Copy code
tracking:
  experiment:
    create_experiment_kwargs: # will be used only if the experiment does not exist yet and is created.    
        artifact_location: null # enable to specify an artifact location for the
Another thing, how could I load parameters from other 
.yml
file into 
mlflow.yml
?
mlflow.yml
is a normal configuration file, so you can use 
${runtime_params: ...}
or 
${globals: ...}
resolvers as usual. See : https://docs.kedro.org/en/stable/configuration/advanced_configuration.html#how-to-use-global-variables-with-the-omegaconfigloader and https://docs.kedro.org/en/stable/configuration/advanced_configuration.html#how-to-override-configuration-with-[…]rameters-with-the-omegaconfigloader
Open in Slack
PreviousNext
Powered by