Module: 03

Lab/Demo: 03

Task: 00

Step: 00

Description of issue - Replace Azure AI Services instead of Cognitive Services in all Labs steps

Run an automated machine learning job
Step 1.

Author, should be "Authoring"

Module: Anomaly Detector

Lab/Demo: 01

Task: 00

Step: 00

Description of issue
Known issue -- The create icon for the anomaly detector in the lab was missing and therefore the students couldn't complete lab 1.

Question:
If the students completed the other labs in the module, will they have acquired the knowledge necessary for AI-900?

Description of issue: I have recently cleared my AI-900 exam and can guide learners with sample questions to help with the exams by compiling data from various resources I used in my preparation.

what is issue when we run the command in of step one in powershell after cloning then they give of error

AI-900-AIFundamentals/instructions/02a-create-regression-model.html

Description of issue: Python incorrect on Create and run an inference pipeline Step 9 is incorrect. Code should have an indented return:

import pandas as pd

def azureml_main(dataframe1=None, dataframe2=None):
    scored_results = dataframe1[['Scored Labels']]
    scored_results.rename(columns={'Scored Labels': 'predicted_price'}, inplace=True)
    return scored_results

I've already left a comment on the commit where the change was done. Tagging: @sherzyang

Module: AI-900

Lab/Demo: 02b

Task: Add transformations

Step: 2 - 3

Description of issue
I think that not there is a step missing like
'add all but pationt ID columns to select columns'
maybe even with images and things.

Double click on the Select Columns in Dataset module to access a settings pane on the right. Select Edit column. Then in the Select columns window, select By name and Add all to add all the columns. Then remove PatientID, so your final column selection looks like this:

Module: Analyze images with the Computer Vision service

Lab/Demo: Exercise - Explore Computer Vision

Task: Configure and run a client application

Step: 6

Description of issue
When running the script in Azure Cloud Shell, received the follow error

PS /home/user1-30332935/ai-900> ./analyze-image.ps1 store-camera-1.jpg
Analyzing image...
Invoke-RestMethod: /home/user1-30332935/ai-900/analyze-image.ps1:21
Line |
  21 |  $result = Invoke-RestMethod -Method Post `
     |            ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     | {"error":{"code":"404","message": "Resource not found"}}
ConvertFrom-Json: /home/user1-30332935/ai-900/analyze-image.ps1:26
Line |
  26 |  $analysis = $result | ConvertFrom-Json
     |                        ~~~~~~~~~~~~~~~~
     | Cannot bind argument to parameter 'InputObject' because it is null.

Description:

Objects in this image:

Tags relevant to this image:

Repro steps:
Follow the instructions in the lab.
PowerShell Version: 7.3.3

Module: Explore clustering with Azure Machine Learning Designer

Exercise: Unit 6

Task: Test the service

Step: 2

Description of issue
the following code-
{
"Inputs": {
"WebServiceInput0": [
{
"CulmenLength": 49.1,
"CulmenDepth": 4.8,
"FlipperLength": 1220,
"BodyMass": 5150
}
]
},
"GlobalParameters": {}
}

-------------gives this error--->
Failed to test real-time endpoint
{"error":{"code":400,"message":"Input Data Error. Input data are inconsistent with schema.\nSchema: {'input1': '1:Dataset'}\nData: {'WebServiceInput0': defaultdict(<class 'list'>, {'CulmenLength': [49.1], 'CulmenDepth': [4.8], 'FlipperLength': [1220], 'BodyMass': [5150]})}\nNoneType: None\n","details":""}}

Hence, the error can be corrected by -->
{
"Inputs": {
"input1": [

  {
    "CulmenLength": 49.1,
    "CulmenDepth": 4.8,
    "FlipperLength": 1220,
    "BodyMass": 5150
  }
  
]

},
"GlobalParameters": {}
}

Module: 02

Lab/Demo: Explore Automated Machine Learning in Azure ML

Task: Deploy a predictive service

Step: 03

Description of issue

Button in the Deploy a model doesn't work/do anything after pressing it while trying to deploy a model. Not the only one encountering issue with this step. https://www.reddit.com/r/AzureCertification/comments/13cy5no/ai900_first_lab_does_not_create_the_azure/

Repro steps:

1. Follow the lab instructions

Module: 02

Lab/Demo: 00

Task: 05

Description of issue
The Datastore and file selection portion of instructions is no longer needed. When creating a Dataset in Azure from web files the user is only required to enter the URL address of that web file for Azure to grab those files.

05-create-cognitive-search-solution

The query below did not return correct results.

search=$filter=locations eq 'Chicago'
search=$filter=sentiment eq 'negative'

The query below returned correct results.

search=locations:'Chicago'
search=sentiment:'negative'

data/ml/daily-bike-share.csv

URL: https://microsoftlearning.github.io/AI-900-AIFundamentals/instructions/02a-create-regression-model.html

Module: Create a pipeline in Designer

Lab/Demo: 00

Task: 00

Step: 2 to 5

Description of issue:
The designer got a new lay-out and step 2 to 5 under 'Create a pipeline in Designer' aren't in line with the current designer anymore. See attachments at the bottom.

Proposition for new steps after current step 1:
2. At the top right-hand side of the screen, select Configure & Submit.
3. In the side bar under step 1 Basics, change the draft name (Pipeline-Created-on-date) to Auto Price Training. Click at the bottom on Next.
4. In step 2 Inputs & Outputs click Next again
5. In step 3 Runtime settings under Select compute type, select Compute cluster. Then under Select Azure ML compute cluster, select the compute cluster you created previously. Click at the bottom on Next.
6. In step 4 Review + Submit click on Submit at the bottom.

Attachments:
create-pipeline-help-current.png -> current image under this step on the URL managed above

create-pipeline-help-new.png -> screenshot of new lay-out

Lab: 01

Task: Create an Anomaly Detector resource

Step: 02

Description of issue

The UI in the Azure portal for Anomaly detector has been updated, which has removed it from availability in create resources [p1], and you can no longer create anomaly detectors on its own blade/page in Azure portal [p2].

[p1]

[p2]
Prior version:

Updated:

Module: Create a regression model with Azure Machine Learning designer

Lab/Demo: Explore Regression with Azure Machine Learning Designer

Task: Create Training Pipeline

Step: 1. Return to the pipeline you created

Description of issue
After running the pipeline which prepares the dataset for model training I am unable to edit the pipeline again without cloning it first - this is if I select the pipeline through the 'Jobs' list under 'Assets' in the menu group on the left.

The documentation should direct users to go to the 'Designer' option under the 'Authoring' group.

Documentation source: instructions/02a-create-regression-model.md

Repro steps:

1. Create pipeline, add transformations, run pipeline.
2. Return to pipeline created in previous step
3. Am now unable to edit pipeline, only option is to clone and make edits to clone.

Module: Recognize and synthesize speech

Lab/Demo: speaking-clock.ps1

Task: 03

Step: 06

Description of issue:

I am getting the "Name or service not known" after filling in the key and URL.

Repro steps:

The README file includes this link Download Latest Student Handbook and AllFiles Content but there is no content once you click it.

Module: Explore Face Recognition

Lab/Demo: create-face-solutions

Task: Configure and run a client application

Step: 6

Description of issue
This module provides instructions to create a Cognitive Services resource for use with a face-detection sample. However, I receive the following error message when running the script:

Line |
  22 |  $result = Invoke-RestMethod -Method Post `
     |            ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     | Face APIs are not supported for Cognitive Service resources created after 06/21/2022"

Module: AI-900-AllFundamentals

Lab/Demo: 02a - Create Regression Model

Task: Add and explore a dataset

Step: 02

Description of issue
Unable to select 'Automobile price data (Raw)' dataset onto the canvas.

Repro steps:
None. This dataset doesn't seem to have been loaded as a Data Asset, in the previous steps.

Module: instructions/02b-create-classification-model.md

Lab/Demo: 00

Task: Create an inference pipeline

Step: 5 and 7

Description of issue

The Image (instructions/media/create-classification-model/inference-changes.png) shows the Modules "Enter Data Manually" and "Web Service Input" to be connected to the "Select Columns in Datasets" Input. Whereas the instruction says to connect them to the "Apply Transformations Module".

Connect the output of the Web Service Input component to the right-side input of the Apply Transformation component that is already on the canvas.

I might be missing something. But i guess it should say: "Connect the output of the Web Service Input component to the right-side input of the Select Columns in Datasets component that is already on the canvas

Regards
Jan

P.S. My first ever GitHub Issue. I hope did it correctly.

Module: week1

Lab/Demo: 1

Task: 1b

Step: 03

Description of issue
getting this error when compiled "speaking-clock.ps1"
13 | $result = Invoke-RestMethod -Method Post `
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| Name or service not known
Repro steps:

Module: AI-900

Lab/Demo: 02b - Explore classification with Azure Machine Learning Designer

Task: Add transformations

Step: 2

Description of issue:

The pipeline does not contain a "Select Columns in Dataset" transformation module, so the model is trained on all columns of the dataset, including "PatientID". This is bad practice, since we know that there is no causal relationship between the ID and the diagnosis. In my tests, the model appears to pick up on a negative correlation between ID and diagnosis: by increasing/decreasing the ID and leaving all other inputs unchanged, one can "flip" the decision of the classifier.

The student should be instructed to add a "Select Columns in Dataset" module to the pipeline (analogous to the previous notebook 02a), which selects all columns except "PatientID".

Repro steps:

1. Follow the instructions in the notebook until you reach the task "Test the service".
2. Use the provided test data to make a prediction. Observe that the prediction is positive ("1") with a probability of ~0.7.
3. In the test data, change the value of "PatientID" from 1882185 to 18821850; leave all other values as they are. Observe that the prediction is now negative ("0") with a probability of ~0.13.

Module: AI-900

Lab/Demo: 02b

Task: Create an inference pipeline:

Step: 05

Description of issue

1. The pipeline does not automatically include a Web Service Input component for models created from custom data sets. Search for a Web Service Input component from the asset library and place it at the top of the pipeline. Connect the output of the Web Service Input component to the right-side input of the Apply Transformation component that is already on the canvas.

should be

1. The pipeline does not automatically include a Web Service Input component for models created from custom data sets. Search for a Web Service Input component from the asset library and place it at the top of the pipeline. Connect the output of the Web Service Input component to input of the Select Columns in Dataset component that is already on the canvas.