ylashin

Connect-PnPOnline –Url "YOUR-SITE_URL" –Credentials (Get-Credential)

$web = Get-PnPWeb 

$files  = Get-PnPListItem -List "YOUR-LIST-NAME"

for($i=0; $i -lt $files.Count; $i++)
{
    $web.Context.Load($files[$i].File)
}

ylashin

function Split-array 
{

  param($inArray,[int]$parts,[int]$size)
  
  if ($parts) {
    $PartSize = [Math]::Ceiling($inArray.count / $parts)
  } 
  if ($size) {
    $PartSize = $size

ylashin

Azure Machine Learning done right!

Disclaimer: this isn't about the old classic drag and drop ML studio!

Developing machine learning models or data science applications on Microsoft stack is not a very nice experience. The main tool is Azure Machine Learning Studio and it doesn't provide a good story for DevOps, Source Control or using open source ML tools/IDEs heavily used by data scientists.

A new cloud service has been recently added to Azure to solve those problems and make ML and app development go hand in hand and in harmony.

Meet Azure Machine Learning Service! It's a one-stop shop that you can use to develop, deploy and manage machine learning models.

ylashin

var webServiceUrl = ConfigurationManager.AppSettings["WebServiceUrl"];
var client = new RestClient(webServiceUrl);

var request = new RestRequest(Method.POST);
request.AddHeader("Content-Type", "application/json");
var payload = PrepareWebRequest(imageFilePath);
request.AddJsonBody(payload);

var response = client.Execute(request);
var content = response.Content;

ylashin

! az login

ylashin

! az account list -o table
! az account set --subscription 'YOUR SUBSCRIPTION NAME'

ylashin

! pip install matplotlib onnx
! wget https://www.cntk.ai/OnnxModels/tiny_yolov2/opset_7/tiny_yolov2.tar.gz
! tar xvzf tiny_yolov2.tar.gz

ylashin

# Check core SDK version number
import azureml.core
print("SDK version:", azureml.core.VERSION)

ylashin

from azureml.core import Workspace
ws = Workspace(subscription_id="YOUR SUBSCRIPTION ID", resource_group="YOLO", workspace_name="YOLO")
print(ws.name, ws.resource_group, ws.location, sep = '\n')

ylashin

model_dir = "tiny_yolov2"
from azureml.core.model import Model
model = Model.register(model_path = model_dir + "//model.onnx",
                       model_name = "tiny_yolov2",
                       tags = {"onnx": "demo"},
                       description = "Tiny YOLO model",
                       workspace = ws)