cyrilthank

Thanks @ustcscgyer I am trying to figure out the best way to store the artifacts on the MLFlow server and the below is not working can you please advise where...

requesting since i start the MLtracking server using mlflow server \ --backend-store-uri file:/mlflowserver_backend_store_uri \ --default-artifact-root ftp://cyril:cyril@localhost/mlflowserver_default_artifact_root \ --host 0.0.0.0 The below mlflowClient.create_run(experiment_to_run.experiment_id) puts files into /mlflowserver_backend_store_uri but artifacts are not...

can you please advise me on what i may be doing wrong since the below code creates the artifacts in the mlruns directory on local server import mlflow import mlflow.pyfunc...

Requesting since i need to use mlflow.pyfunc.log_model instead of mlflow.sklearn.log_model(lr, "model")

Here is what i did step 1 : lflow server \ --backend-store-uri file:/mlflowserver_backend_store_uri \ --default-artifact-root file:/mlflowserver_default_artifact_root \ --host 0.0.0.0 Step 2: import os import mlflow import mlflow.pyfunc class AddN(mlflow.pyfunc.PythonModel): def...

Thanks @sid for your response I too am waiting for some clarity on the MLFLOW_TRACKING_URI not sure how folks are using this in production/poc

@wjarek2 can you please advise if tf.estimators you referred to in https://github.com/tensorflow/tfx/issues/205#issuecomment-499252450 can be used in 'some way' to 'bring into TFX flow' the custom pkl file?

Thanks Charles for your patient reply  I still would like to recommend tfx to customersSo can you please share any way such customers can be on boarded Can you please...

@rmothukuru Thanks for your reply However, it is precisely over this feature we abandoned TFX in favor of kubeflow. Thank you cyril

sorry @shazamkash i got 'famously' stuck at https://github.com/kubeflow/kubeflow/issues/4228 https://github.com/kubeflow/kubeflow/issues/4184 both have been marked duplicate So i take it i have to wait for kubeflow 0.7 version to be released and...