Skip to content

Commit 7a70514

Browse files
authored
Update azure-machine-learning-release-notes.md
1 parent 4c7f921 commit 7a70514

File tree

1 file changed

+47
-0
lines changed

1 file changed

+47
-0
lines changed

articles/machine-learning/azure-machine-learning-release-notes.md

Lines changed: 47 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,53 @@ In this article, learn about Azure Machine Learning releases. For the full SDK
1818

1919
See [the list of known issues](resource-known-issues.md) to learn about known bugs and workarounds.
2020

21+
## 2020-03-02
22+
23+
### Azure Machine Learning SDK for Python v1.1.2rc0
24+
25+
+ **New features**
26+
+ [Insert new features below. Reference articles and/or doc pages]
27+
28+
+ **Preview features**
29+
+ [Contrib features below]
30+
31+
+ **Breaking changes**
32+
+ [Reference upcoming breaking changes and old API support drop date]
33+
34+
+ **Bug fixes and improvements**
35+
+ **azureml-automl-core**
36+
+ Enabled the Batch mode inference (taking multiple rows once) for automl ONNX models
37+
+ Improved the detection of frequency on the data sets, lacking data or containing irregular data points
38+
+ Added the ability to remove data points not complying with the dominant frequrncy.
39+
+ **azureml-automl-runtime**
40+
+ Fixed the issue with the error thrown if the grain which was not present in the training set appeared in the test set
41+
+ Removed the y_query requirement during scoring on forecasting service
42+
+ **azureml-contrib-mir**
43+
+ Adds functionality in the MirWebservice class to retrieve the Access Token
44+
+ **azureml-core**
45+
+ Deployed AzureML Webservices will now default to `INFO` logging. This can be controlled by setting the `AZUREML_LOG_LEVEL` environment variable in the deployed service.
46+
+ Fix iterating on `Dataset.get_all`.
47+
+ Improve error message when invalid type is passed to `path` argument of dataset creation APIs.
48+
+ Python sdk uses discovery service to use 'api' endpoint instead of 'pipelines'.
49+
+ Swap to the new routes in all SDK calls
50+
+ Changes routing of calls to the ModelManagementService to a new unified structure
51+
+ Made workspace update method publicly available.
52+
+ Added image_build_compute parameter in workspace update method to allow user updating the compute for image build
53+
+ Added deprecation messages to the old profiling workflow. Fixed profiling cpu and memory limits
54+
+ **azureml-interpret**
55+
+ update azureml-interpret to interpret-community 0.6.*
56+
+ **azureml-mlflow**
57+
+ Add support for sovereign clouds to azureml.mlflow
58+
+ **azureml-pipeline-steps**
59+
+ Moved the `AutoMLStep` to the `azureml-pipeline-steps package`. Deprecated the `AutoMLStep` within `azureml-train-automl-runtime`.
60+
+ **azureml-train-automl-client**
61+
+ Fixed an issue where certain packages may be installed at incorrect versions on remote runs.
62+
+ **azureml-train-automl-runtime**
63+
+ Fixed the issue with frequency detection in the remote runs
64+
+ Moved the `AutoMLStep` to the `azureml-pipeline-steps package`. Deprecated the `AutoMLStep` within `azureml-train-automl-runtime`.
65+
+ **azureml-train-core**
66+
+ Moved the `AutoMLStep` to the `azureml-pipeline-steps package`. Deprecated the `AutoMLStep` within `azureml-train-automl-runtime`.
67+
2168
## 2020-02-18
2269

2370
### Azure Machine Learning SDK for Python v1.1.1rc0

0 commit comments

Comments
 (0)