CODEWITHSUNDEEP

scikit-learndatabricksmlflow

Jaroslav Bezděk

Jaroslav Bezděk

Reputation: 7635

MlflowException: RestException('BAD_REQUEST: CreateMultipartUpload with multi-cloud URLs is supported for AWS only.')

I try to train and then log my scikit-learn model into Databricks model registry. I run the following code on Databricks serverless cluster.

mlflow.sklearn.log_model(
    sk_model=model,
    artifact_path="model",
    registered_model_name=model_name,
    signature=infer_signature(X_train, y_train),
    input_example=X_train.head(),
)

Sometimes the code finishes without problems, but sometimes it throws the following error:

MlflowException: The following failures occurred while uploading one or more artifacts to dbfs:/databricks/mlflow-tracking/3584249696627184/1eb4b9879a8f4f86a5bfb2a79bb078d1/artifacts: {'/local_disk0/repl_tmp_data/ReplId-1953d-b4049-f/tmpr0nr43xl/model/model.pkl': "RestException('BAD_REQUEST: CreateMultipartUpload with multi-cloud URLs is supported for AWS only.')"}

What am I doing wrong?

Upvotes: 0

Views: 15

Answers (0)

Related Questions