Skip to content

Commit 8e73c60

Browse files
juliakmjuliakm
committedOct 19, 2020
Adding GitHub Actions for Machine Learning article
1 parent a8939c9 commit 8e73c60

File tree

3 files changed

+224
-1
lines changed

3 files changed

+224
-1
lines changed
 

‎articles/app-service/deploy-github-actions.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -119,7 +119,7 @@ When you configure your GitHub workflow, you use the `AZURE_WEBAPP_PUBLISH_PROFI
119119
120120
In [GitHub](https://github.com/), browse your repository, select **Settings > Secrets > Add a new secret**.
121121
122-
To use [user-level credentials](#generate-deployment-credentials), paste the entire JSON output from the Azure CLI command into the secret's value field. Give the secret the name like `AZURE_CREDENTIALS`.
122+
To use [user-level credentials](#generate-deployment-credentials), paste the entire JSON output from the Azure CLI command into the secret's value field. Give the secret the name `AZURE_CREDENTIALS`.
123123

124124
When you configure the workflow file later, you use the secret for the input `creds` of the Azure Login action. For example:
125125

‎articles/machine-learning/how-to-github-actions-machine-learning.md

Lines changed: 223 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,223 @@
1+
---
2+
title: GitHub Actions for CI/CD
3+
titleSuffix: Azure Machine Learning
4+
description: Learn about how to create a GitHub Actions workflow to train a model on Azure Machine Learning
5+
services: machine-learning
6+
ms.service: machine-learning
7+
ms.subservice: core
8+
author: juliamkm
9+
ms.author: jukullam
10+
ms.date: 10/19/2020
11+
ms.topic: conceptual
12+
ms.custom: github-actions-azure
13+
---
14+
15+
# Deploy to App Service using GitHub Actions
16+
17+
Get started with [GitHub Actions](https://help.github.com/en/articles/about-github-actions) to train a model on Azure Machine Learning.
18+
19+
## Prerequisites
20+
21+
- An Azure account with an active subscription. [Create an account for free](https://azure.microsoft.com/free/?WT.mc_id=A261C142F).
22+
- A GitHub account. If you don't have one, sign up for [free](https://github.com/join).
23+
24+
## Create repository
25+
26+
Create a new repository off the [ML Ops with GitHub Actions and Azure Machine Learning template](https://github.com/machine-learning-apps/ml-template-azure).
27+
28+
1. Open the [template](https://github.com/machine-learning-apps/ml-template-azure) on GitHub.
29+
2. Select **Use this template**.
30+
31+
:::image type="content" source="media/how-to-github-actions-machine-learning/gh-actions-use-template.png" alt-text="Select use this template":::
32+
3. Create a new repository from the template. Set the repository name to `ml-learning` or a name of your choice.
33+
34+
35+
## Workflow file overview
36+
37+
A workflow is defined by a YAML (.yml) file in the `/.github/workflows/` path in your repository. This definition contains the various steps and parameters that make up the workflow.
38+
39+
The file has four sections:
40+
41+
|Section |Tasks |
42+
|---------|---------|
43+
|**Authentication** | 1. Define a service principal. <br /> 2. Create a GitHub secret. |
44+
|**Connect** | 1. Connect to the machine learning workspace. <br /> 2. Connect to a compute target. |
45+
|**Run** | 1. Submit a training run. |
46+
|**Deploy** | 1. Register model in Azure Machine Learning registry. 1. Deploy the model. |
47+
48+
## Generate deployment credentials
49+
50+
You can create a [service principal](../active-directory/develop/app-objects-and-service-principals.md#service-principal-object) with the [az ad sp create-for-rbac](/cli/azure/ad/sp?view=azure-cli-latest#az-ad-sp-create-for-rbac&preserve-view=true) command in the [Azure CLI](/cli/azure/). Run this command with [Azure Cloud Shell](https://shell.azure.com/) in the Azure portal or by selecting the **Try it** button.
51+
52+
```azurecli-interactive
53+
az ad sp create-for-rbac --name "myML" --role contributor \
54+
--scopes /subscriptions/<subscription-id>/resourceGroups/<group-name> \
55+
--sdk-auth
56+
```
57+
58+
In the example above, replace the placeholders with your subscription ID, resource group name, and app name. The output is a JSON object with the role assignment credentials that provide access to your App Service app similar to below. Copy this JSON object for later.
59+
60+
```output
61+
{
62+
"clientId": "<GUID>",
63+
"clientSecret": "<GUID>",
64+
"subscriptionId": "<GUID>",
65+
"tenantId": "<GUID>",
66+
(...)
67+
}
68+
```
69+
70+
## Configure the GitHub secret
71+
72+
1. In [GitHub](https://github.com/), browse your repository, select **Settings > Secrets > Add a new secret**.
73+
74+
2. Paste the entire JSON output from the Azure CLI command into the secret's value field. Give the secret the name `AZURE_CREDENTIALS`.
75+
76+
## Connect to the workspace
77+
78+
Use the `Azure/aml-workspace@v1` action to connect to your Azure Machine Learning workspace.
79+
80+
```yaml
81+
- name: Connect/Create Azure Machine Learning Workspace
82+
id: aml_workspace
83+
uses: Azure/aml-workspace@v1
84+
with:
85+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
86+
```
87+
88+
By default, the action expects a `workspace.json` file. If your JSON file has a different name, you can specify it with the `parameters_file` input parameter. If there is not a file, a new one will be created with the repository name.
89+
90+
91+
```yaml
92+
- name: Connect/Create Azure Machine Learning Workspace
93+
id: aml_workspace
94+
uses: Azure/aml-workspace@v1
95+
with:
96+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
97+
parameters_file: "alternate_workspace.json"
98+
```
99+
The action writes the workspace Azure Resource Manager (ARM) properties to a config file, which will be picked by all future Azure Machine Learning GitHub Actions. The file is saved to `GITHUB_WORKSPACE/aml_arm_config.json`.
100+
101+
## Connect to a Compute Target in Azure Machine Learning
102+
103+
Use the `Azure/aml-compute@v1` action to connect to a compute target in Azure Machine Learning. If the compute target exists, the action will connect to it. Otherwise the action will create a new compute target. The AML Compute action only supports Azure ML Clusters and AKS Clusters.
104+
105+
```yaml
106+
- name: Connect/Create Azure Machine Learning Compute Target
107+
id: aml_compute_training
108+
uses: Azure/aml-compute@v1
109+
with:
110+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
111+
```
112+
## Submit training run
113+
114+
Use the `Azure/aml-run@v1` action to submit a ScriptRun, an Estimator or a Pipeline to Azure Machine Learning.
115+
116+
```yaml
117+
- name: Submit training run
118+
id: aml_run
119+
uses: Azure/aml-run@v1
120+
with:
121+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
122+
```
123+
124+
## Register model in registry
125+
126+
Use the `Azure/aml-registermodel@v1` action to register a model to Azure Machine Learning.
127+
128+
```yaml
129+
- name: Register model
130+
id: aml_registermodel
131+
uses: Azure/aml-registermodel@v1
132+
with:
133+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
134+
run_id: ${{ steps.aml_run.outputs.run_id }}
135+
experiment_name: ${{ steps.aml_run.outputs.experiment_name }}
136+
```
137+
138+
## Deploy model to Azure Machine Learning to ACI
139+
140+
Use the `Azure/aml-deploy@v1` action to deploys a model and create an endpoint for the model.
141+
142+
```yaml
143+
- name: Deploy model
144+
id: aml_deploy
145+
uses: Azure/aml-deploy@v1
146+
with:
147+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
148+
model_name: ${{ steps.aml_registermodel.outputs.model_name }}
149+
model_version: ${{ steps.aml_registermodel.outputs.model_version }}
150+
151+
```
152+
153+
## Complete example
154+
155+
Train your model and deploy to Azure Machine Learning.
156+
157+
```yaml
158+
# Actions train a model on Azure Machine Learning
159+
name: Azure Machine Learning training and deployment
160+
on:
161+
push:
162+
branches:
163+
- master
164+
# paths:
165+
# - 'code/*'
166+
jobs:
167+
train:
168+
runs-on: ubuntu-latest
169+
steps:
170+
# Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
171+
- name: Check Out Repository
172+
id: checkout_repository
173+
uses: actions/checkout@v2
174+
175+
# Connect or Create the Azure Machine Learning Workspace
176+
- name: Connect/Create Azure Machine Learning Workspace
177+
id: aml_workspace
178+
uses: Azure/aml-workspace@v1
179+
with:
180+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
181+
182+
# Connect or Create a Compute Target in Azure Machine Learning
183+
- name: Connect/Create Azure Machine Learning Compute Target
184+
id: aml_compute_training
185+
uses: Azure/aml-compute@v1
186+
with:
187+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
188+
189+
# Submit a training run to the Azure Machine Learning
190+
- name: Submit training run
191+
id: aml_run
192+
uses: Azure/aml-run@v1
193+
with:
194+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
195+
196+
# Register model in Azure Machine Learning model registry
197+
- name: Register model
198+
id: aml_registermodel
199+
uses: Azure/aml-registermodel@v1
200+
with:
201+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
202+
run_id: ${{ steps.aml_run.outputs.run_id }}
203+
experiment_name: ${{ steps.aml_run.outputs.experiment_name }}
204+
205+
# Deploy model in Azure Machine Learning to ACI
206+
- name: Deploy model
207+
id: aml_deploy
208+
uses: Azure/aml-deploy@v1
209+
with:
210+
azure_credentials: ${{ secrets.AZURE_CREDENTIALS }}
211+
model_name: ${{ steps.aml_registermodel.outputs.model_name }}
212+
model_version: ${{ steps.aml_registermodel.outputs.model_version }}
213+
214+
```
215+
216+
## Clean up resources
217+
218+
When your resource group and repository are no longer needed, clean up the resources you deployed by deleting the resource group and your GitHub repository.
219+
220+
## Next steps
221+
222+
> [!div class="nextstepaction"]
223+
> [Create and run machine learning pipelines with Azure Machine Learning SDK](how-to-create-your-first-pipeline.md)

0 commit comments

Comments
 (0)
Please sign in to comment.