Open jknmsft opened 2 years ago
I had the same problem and I found out that it was because the Databricks runtime 11.0 ML is using Python 3.9. But azureml-sdk is only working for Python versions up to 3.8. So, I created a new cluster with Databricks Runtime 10.4 ML (Python 3.8) and that did the trick
I have tried installing azureml-sdk[databricks] on a Azure Databricks cluster, runtime 11.0 ML (includes Apache Spark 3.3.0, Scala 2.12), node type Standard_DS_v2 in two different ways 1) Using the UI following the instructions in the Track Azure Databricks experiments in Azure Machine Learning course (https://docs.microsoft.com/en-us/learn/modules/track-azure-databricks-experiments-azure-machine-learning/?ns-enrollment-type=learningpath&ns-enrollment-id=learn.wwl.build-operate-machine-learning-solutions-azure-databricks) and 2) With !pip install azureml-sdk[databricks]
With (1) installation never completes With (2) installation fails. The output shows several instances of INFO: pip is looking at multiple versions of azureml-train to determine which version is compatible with other requirements. This could take a while. and INFO: This is taking longer than usual. You might need to provide the dependency resolver with stricter constraints to reduce runtime. See https://pip.pypa.io/warnings/backtracking for guidance. If you want to abort this run, press Ctrl + C.
After about 90 seconds installation fails with the following error message