Updates to Azure Arc-enabled Machine Learning

by Contributed | Jun 21, 2021 | Technology

This article is contributed. See the original author and article here.

Azure Machine Learning (AML) team is excited to announce the availability of Azure Arc-enabled Machine Learning (ML) public preview release. All customers of Azure Arc-enabled Kubernetes now can deploy AzureML extension release and bring AML to , and the edge using Kubernetes on their hardware of choice.

The design for Azure Arc-enabled ML helps IT Operators leverage native Kubernetes concepts such as namespace, node selector, and resources requests/limits for ML compute utilization and optimization. By letting the IT operator manage ML compute setup, Azure Arc-enabled ML creates a seamless AML experience for data scientists who do not need to learn or use Kubernetes directly. Data scientists now can focus on models and work with tools such as Azure Machine Learning AML Studio, AML 2.0 CLI, AML Python SDK, productivity tools like Jupyter notebook, and ML frameworks like TensorFlow and PyTorch.

IT Operator experience – ML compute setup

Once Kubernetes cluster is up and running, IT Operator can follow 3 simple steps below to prepare cluster for AML workload:

• Connect Kubernetes cluster to Azure via Azure Arc


• Deploy AzureML extension to Azure Arc-enabled cluster


• Create a compute target for Data Scientists to use

For the first two steps, IT Operator can simply run the following two CLI commands to accomplish:

Once AzureML extension installation completes in cluster, you will see following AzureML pods running inside the :

With your cluster ready to take AML workload, now you can head over to AML Studio portal and create compute target for Data Scientists to use, see below AML Studio compute attach UI:

Note by clicking “New->Kubernetes (preview)”, the Azure Arc-enabled Kubernetes clusters will automatically appear in a dropdown list for IT Operator to attach. In the process of Studio UI attach operation, IT Operator could provide an optional JSON configuration file specifying namespace, node selector, and resources requests/limits to be used for the compute target being created. With these advanced configurations in compute target, IT Operator help Data Scientists target a subset of nodes such as GPU pool or CPU pool for training job, improves compute resource utilization and avoids fragmentation. For more information about creating compute targets using these custom properties, please refer to AML documentation.

For upcoming Azure Arc-enabled ML update release, we plan to support compute target creation through CLI command as well, and ML compute setup experience will be simplified to following 3 CLI commands:

Note when connecting Kubernetes cluster to Azure via Azure Arc, IT Operator can also specify configuration setting to enable outbound proxy server. We are pleased to announce that Azure Arc-enabled Machine Learning fully supports model training on-premises with outbound proxy server connection.

Data Scientist experience – train models

Once the attached Kubernetes compute target is available, Data Scientist can discover the list of compute targets in AML Studio UI compute section. Data Scientist can choose a suitable compute target to submit training job, such as GPU compute target, or CPU compute target with proper resources requests for particular training job workload, such as the # of vCPU cores and memory. Data Scientist can submit job either through AML 2.0 CLI or AML Python SDK, in either case Data Scientist will specify compute target name at job submission time. Azure Arc-enabled ML supports the following built-in AML training features seamlessly:

• Train models with AML 2.0 CLI
  
  
  
  • Basic Python training job
  
  
  • Distributed training – PyTorch
  
  
  • Distributed training – TensorFlow
  
  
  • Distributed training – MPI
  
  
  • Sweep hyperparameters
  
  
  
  


• Train models with AML Python SDK
  
  
  
  • Configure and submit training run
  
  
  • Tune hyperparameters
  
  
  • Scikit-learn
  
  
  • TensorFlow
  
  
  • PyTorch
  
  
  
  


• Build and use ML pipelines including AML Designer pipeline support

For those Data Science professionals who have used AML Python SDK, existing AML Python SDK examples and notebooks, or your existing projects will work out-of-box with a simple change of compute target name in Python script. If you are not familiar with Python SDK yet, please refer to above links to get started.

AML team is extremely excited that Azure Arc-enabled ML supports the latest and greatest AML 2.0 CLI training job submission, which is in public preview also. Train models with AML 2.0 CLI is simple and easy with following CLI command:

Let’s take a look at job YAML file:

Note job YAML file specifies all training job needed resources and assets including training scripts and compute target. In this case, Data Scientist is using Azure Arc-enabled compute target created by IT Operator earlier. Running the job creation CLI command will submit job to Azure Arc-enabled Kubernetes cluster and opens AML Studio UI portal for Data Scientist to monitor job running status, analyze metrics, and examine logs. Please refer to Train models with the 2.0 CLI for more information and examples.

Get started today

In this post, we provided status updates to Azure Arc-enabled Machine Learning and showed how IT Operator can easily setup and prepare Azure Arc-enabled Kubernetes cluster for AML workload, and how Data Scientist can easily train models with AML 2.0 CLI and Kubernetes compute target.

To get started with Azure Arc-enabled ML for training public preview, visit Azure Arc-enabled ML Training Public Preview Github repository, where you can find detailed documentation for IT Operator and Data Scientist, and examples for you to try out easily. In addition, visit the official AML documentation to find more information.

Azure Arc-enabled ML also supports model training with interactive job experience and debugging, which is in private preview. Please sign up here for interactive job private preview.

After Data Scientist trains a model, ML Engineer or Model Deployment Pro can deploy the model with Azure Arc-enabled ML on the same Arc-enabled Kubernetes cluster, which is in private preview too. Please sign up here for inference private preview.

Also, check out these additional great AML blog posts!

• Announcing the new CLI and ARM REST APIs for Azure Machine Learning


• Run Azure Machine Learning anywhere – on hybrid and in multi-cloud with Azure Arc

Save an email attachment to a custom path with Power Automate

by Contributed | Jun 21, 2021 | Technology

This article is contributed. See the original author and article here.

This use case was generated from an idea involving Xerox Docucentres, which enables end-users to email themselves a scanned document. Equally, this applies to regular emails with attachments that you would like Power Automate to automatically save the attachment(s) to a specific (custom/bespoke/unique) folder and/or SharePoint site and therefore Document Library.   The dynamically chosen path is based on the Subject Line OR File Name. This can be achieved relatively easily and would save you a lot of routine administration time.

The Solution

Using Power Automate, When a new email arrives (v3) trigger, we can be quite specific about the sender of the email and for this, I have assumed all emails will come from xerox@mydomain.com and the email must contain an attachment! Ideally, you want to limit the emails that this Flow will trigger by using the available parameters for that action.

 

To define the locations for saving files, I have used an array in a compose action (but don’t panic!). You might want to use a SharePoint list as somewhere to save the specific subject or file name keywords, alongside the folder or SharePoint site path. The advantage of using an array is that only the owner(s) of the flow can define these and all of the parameters are in one place. Creating a list in SharePoint might give you and/or other colleagues the ability to easily define and visualise the locations.

 

The parameters for this flow are basic, a keyword for searching the Subject or Filename containing, which I have called Dept; and the Path, which can either be a folder or a Site Name. If you are looking to have a dynamic Folder and Site, you can define a third parament or key:value pair.

Finding our path

We must now use the filter array action, which will allow us to search the Subject or Attachment name to see if it contains one of the keywords from the Dept key. If we get a match, the array will return a value for the Path. Here I use a compose action to return the first object from the array result. Why do I use the first expression? Because the result is returned as an array, it expects multiple elements.  Calling the path without either using the first expression or body(‘FilterFindPath’)?[0]?[‘Path’] to select the first element, you will receive an error “Array elements can only be selected using an integer index“. Using First() or calling the element by using an integer will allow you to simplify your Flow if you know the result will always be the first element.

 

Saving the File(s) to a Custom Path

Using the Create File action for SharePoint, I am using the Attachments Name and Attachments Content dynamic expressions from the email trigger. By default, when you select these dynamic values, Power Automate will put your actions into an Apply to Each. Why? This is because the results for the attachment name and content are returned as an array. You can accept this and your solution will handle multiple attachments. For my solution, I am a stickler for efficiency and have used the first() expression described previously as my Xerox will only ever include one file. You can accept the default Power Automate build if you would prefer.

I’ve two actions here, one to demonstrate how to save to a custom dynamic path, the other to a custom dynamic site. When manually supplying a custom path for the site or folder, you will have to select Enter Custom Value or Power Automate will try and resolve the address and error ‘Site Address’ is required or the folder path with try and delete any additional text that you supply.

 

The expressions I have used both using the integer selection for the first array element and the first expression are as follows:

File Name: triggerOutputs()?[‘body/attachments’]?[0]?[‘name’]
File Content: first(triggerOutputs()?[‘body/attachments’])?[‘contentBytes’]

To finish off my flow and this is personal preference, I have deleted the incoming email. Why keep it? The file attachment has been saved for me.

 

See it in Action

Here I have an incoming email from the Xerox mailbox. I’ve been sent me an attachment from the Xerox Mailbox. This will trigger my flow.

The filter action will look for each of the Dept values in the subject line and return a filtered array. In the example below, we output the ICT Path based on the subject line “This file is for ICT thanks”.

 

What do the two Create File Actions look like? Here you can see the custom dynamic paths for both the Folder Path or Site Address. I am obviously saving the file twice here in order to demonstrate the two options. You might want to combine the custom site and folder with an additional parameter in your Array or SharePoint List.

 

Summary

Quickly file away those routine attached documents received via email. Maybe this would simplify uploading documents into SharePoint? Work on a file and email it to a shared mailbox and have the Flow decide where to file the attachment based on keywords? Please let me know how you could use this.

Make sure you check out my YouTube channel for other ideas and proof of concepts.

How to capture exception and dump files of LogicApp standard in Kudu console

by Contributed | Jun 20, 2021 | Technology

This article is contributed. See the original author and article here.

The tool:

SmartDump – an exception and memory dump capture utility (github.com)     –              https://github.com/microsoft/SmartDump

Current Release v1.02 beta · microsoft/SmartDump

SmartDump_v1.02.zip

How to use:

1. First of all, we need to open Kudu console and drag/upload the tool into the site folder.

2. Next, we need to find the PID of our LogicApp site’s w3wp.exe from Process Explorer.

NOTE: the one with (scm) is for Kudu console. Hence we always have to focus on w3wp.exe without (scm).

3. Then we can run SmartDump.exe and use -p option to specify the PID found in step #2.

This makes SmartDump attach to the target process and start to monitor any exception thrown in it. By default, the tool captures 5 exceptions.

After invoke LogicApp and generate some exceptions, SmartDump will be able to capture them within the console.

4. If you want to see more exceptions, just use -n option to specify number of exceptions to be captured.

NOTE: using -n 0 to start an unlimited/endless capture. However please be careful to use this in Kudu because it doesn’t support Ctrl+C to exit a process. Run inside a common cmd.exe without such issue.

Tips: you can also add:  > filename.log  at the end of a command to make Kudu write output into a log file for you.

5. To generate dump, use -d option to set number of dumps to be captured. Associating it with -f(filter include) and -fv (filter exclude) options allow you to capture dumps against specific exceptions.

-f      Filter exception based on specified string(s). Use ‘|’ as delimiter for multiple strings.

-fv     Exclude exceptions contain specified filter. Use ‘|’ as delimiter for multiple strings.

6. The tool also supports to set memory address of breakpoint to generate dump files.

You can capture a dump first and then open it in debugger to find the code entry address of a function you interest in(or code address of any line).

Then use -a option of SmartDump to set the address as breakpoint for capture:

There are also some other useful options and sample commands can be found in the readme of GitHub page: SmartDump

Upcoming June 2021 Microsoft 365 Champion Community Call

by Contributed | Jun 18, 2021 | Technology

This article is contributed. See the original author and article here.

Join us for this month’s community call where we will continue with our every 4th Tuesday of the month schedule, occurring on June 22nd! Join us at either 8:00 AM or 5:00 PM PT.

This month we will be discussing Lists and To-Do as our primary topics. We will review new features for Lists and To-Do and share tips that you can use to inspire new ways for your organizations to leverage these solutions.

If you have not yet joined our champion community, signup here to get the resource links that contain access to the call calendar, invites, program assets, and previous calls!

http://aka.ms/m365champions 

We look forward to seeing you there!

/Josh

[Amplifying Black Voices] Availability Vs. Accessibility: Creating Seats At a Hidden Table

by Contributed | Jun 18, 2021 | Technology

This article is contributed. See the original author and article here.

This blog was written by Program Manager, Azure Fastrack, iAsia Brown as part of the Amplifying Black Voices blog series. iAsia reflects on Juneteenth and how it correlates to her efforts to introduce and usher people into the field of technology.

On Wednesday, 16 June 2021, the Senate passed a bill to establish Juneteenth as a national holiday, and with mixed emotions, I celebrated. I celebrated the final acknowledgment that there was a period when slaves, black people, were deliberately kept in the dark for three years that they were free. Imagine gatekeeping a person’s ability to be more than property, live their lives on their terms, and find adequate employment.

After 16 years in the military, I remember when I first got my DD-214, my discharge papers, the papers that say I’ve completed serving the Marine Corps. I was free to live this life as a civilian and do whatever it was civilians do. I had been in the military since I was 17. It was my first and only real job. I didn’t know what I wanted to do. I remember me declaring, ” I will get out of the military and go work at Microsoft.” Once I got here, my immediate reaction was, “now that I have access, who else can I bring along?”

On Day 2 of my new life, I posted a LinkedIn message about my new employment and willingness to help anyone looking for help pivoting into tech. I can say that within two years, I’ve helped over 102 different humans across the US find meaningful work at various tech companies to include, but not limited to, Microsoft and other big tech companies. Anyone who asked, I was happy to help. I was more than creating seats at hidden tables. Some tables are hidden by design, others due to lack of access to someone with the knowledge, active with access in these spaces, and willing to help them on their journey there.

If Juneteenth has taught me anything in this lifetime, it’s that not everyone will pass along the information to allow you access to new opportunities. When you find someone willing to help, take it and then pay it forward. I want to help be a catalyst to the change I want to see in tech. I want to see more women, more BIPOC, more LQBTQIA+, more veterans, more people who may be disabled but able and wanting to do these jobs. I want everyone to win. 

Technology is the equalizer. It doesn’t matter your race, religion, sexual orientation, or height. None of it has a discriminatory place here, yet all of the beautiful things that make people different are needed here. Let’s continue to be the representation the generation coming behind us can see; let’s continue to make space at tables that we are discovering and uncovering for those we are bringing along. As we continue to pay it forward  and reflect on this holiday, I will challenge myself to continue to provide access to communities who were raised believing that it was not available.