- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-09-2022 04:20 AM
I have tried to set up a autoML experiment with runtime 11.2ML and data from a delta table. However I receive the error "ModuleNotFoundError: No module named 'databricks.automl'" and "AutoML not available: Use Databricks Runtime 8.3 ML or above." though I am using a runtime newer than 8.3. Does anyone know how to get the module so I can proceed with autoML?
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-13-2022 03:54 AM
I deleted the cluster and created a new with runtime 9.1 LTS ML which solved the problem.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-09-2022 02:44 PM
It should be
from databricks import automl
The table doesn't need to be delta, but does need to be a table in the metastore.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-11-2022 11:15 PM
Hi @Joseph Kambourakis
When I run that code I get the message on the figure below. It seems that my "databricks" library doesn't include automl.
(I am using runtime 11.2 ML though it states that autoML is not available for my cluster. I have tried earlier versions too.)
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-12-2022 09:29 AM
You do not need to install this library. It is automatically installed within the newest (10.0 or above) Databricks ML runtimes. Use 10.4 ML LTS for the more stable runtime and use the newest runtimes for the latest features.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-12-2022 11:18 PM
hi @Salma Mayorquin
In the original post I was using runtime 11.2ML and I have now tried with runtime 10.4ML, but I still get the same error.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-19-2022 03:20 PM
Just to confirm, this is correct, and I do not see this problem on 11.2 ML. The software is installed, and your import is correct. I am not sure why you observe this; are init scripts modifying the ML runtime when the cluster starts, maybe?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-19-2022 11:03 PM
I used the script provided by databricks as I use the autoML incorporated in databricks, so I would not assume that it modifies the ML runtime. I think there must have been something wrong with the cluster configuration, because I have created a new cluster and now it seems to work.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
09-13-2022 03:54 AM
I deleted the cluster and created a new with runtime 9.1 LTS ML which solved the problem.