DP-100 Azure Machine Learning in Python-Basic to Advance2023

Create and manage a local virtual environment for azure machine learning using conda, anaconda, and pip, install libraries from an environment file, and run code locally or in the cloud.

Train a model in a virtual environment using a diabetes dataset with a random forest classifier, evaluate with accuracy and a classification report, and save the trained model with joblib.

Explore workspace capabilities in Azure Machine Learning, including workspace creation, launch studio access, notebooks, automated machine learning, and designer for training, evaluating, and deploying models.

Submit training code to an Azure machine learning workspace, configure a compute target and data storage, create and run an experiment, and troubleshoot errors using logs.

Set up an Azure ML environment with the required libraries, run a training script inside Azure ML, and monitor experiments with logs and snapshots to achieve about 77 percent accuracy.

Master passing runtime arguments in Python to tune a random forest classifier without hard coding, using argparse, and observe how different parameters affect accuracy on local and Asian machines.

Log metrics with the Azure Machine Learning Python SDK, capturing parameters and accuracy from front-end runs. Compare runs to identify the best model by tuning parameters like minimum leaves.

Deploy a machine learning model in Azure ML by preparing the scoring and prediction script, creating an inference environment, configuring deployment settings, and publishing an endpoint for predictions.

Deploy Azure ML models by building a conda-based environment with required libraries, adding dependencies via pip, and exporting a serialized environment file for seamless model deployment.

Create a yaml file to build a virtual environment for deep learning, specifying python 3.8.5, anaconda libraries, and Ramell SDK, then save and run to create and activate the environment.

learn to build a pytorch deep learning model by preparing data, converting features and labels to tensors, creating a dataset and dataloader, and training a neural network with sigmoid activation.

Learn to train a PyTorch model in Azure ML using Python SDK by creating an environment from a spec file, wrapping code in a script config, and submitting the experiment.

Learn to connect an external Azure SQL Database to Azure ML, create a data frame from the dataset, and train and score models via scripts and experiments.

Set up an Azure machine learning pipeline in Python by creating two clusters for preparation and training, then configure steps, environments, and an experiment to run the pipeline.

Learn to schedule an Azure Machine Learning pipeline with a recurrence object, setting frequency (daily, monthly) and intervals, and configure workspace, pipeline name, description, and experiment name for automated runs.

Azure AutoML automatically creates datasets from local data, registers them, runs multiple classification algorithms, and identifies the best model for the diabetes dataset.