• Feature
• The file structure in the Github repository
• Azure Cloud service used
• Pipeline Jobs Metrics
  • Jobs
  • Metrics
• Models
• Blob
• License

• Send the CSV files from the edge to Azure Blob.
• Connect AzureML Datastore to Azure Blob.
• Read CSV data from AzureML Datastore and create corresponding Dataset object.
• Create Azure ML PipelineData objects to store intermediate data generated in the training and evaluation steps of the pipeline.
• Create two steps, Training Step and Evaluate Step, through Azure ML Pipeline. The Training Step is responsible for training the model and storing the generated model in the Azure ML Models Assets and Azure Blob. The Evaluate Step is responsible for evaluating the performance of the model using test data.

├── AzureBlob
│   ├── ListDeleteFiles
│   │   ├── app.py
│   │   └── requirements.txt
│   └── UploadFiles
│       ├── app.py
│       └── requirements.txt
├── AzureML
│   ├── pipeline-python
│   │   ├── evaluate.py
│   │   └── training.py
│   └── run_iris_pipeline.ipynb
├── Datasets
│   ├── test
│   │   └── iris_test.csv
│   └── training
│       └── iris_training.csv
├── LICENSE
└── README.md

• Azure Machine Learning Studio
• Azure VM
• Azure Container Registry
• Azure Storage Account

• Jobs ( ① Training Step → ② Evaluate Step )

• Metrics

  • Training Step
  • Evaluate Step

Examples of AzureML is licensed under the MIT license.