Deployment of Machine Learning models in Azure machine learning through mlflow

[Likith]

MLOps are new to me. If someone could describe the steps for using mlflow in the deployment of Machine Learning Certification Training models in Azure Machine Learning, that would be great. I'm not interested in using databricks. A sample run through each stage with an example would be beneficial. Thank you in advance.

[PadmaPriya · Accepted solution]

Hi @Likith 

Thanks for posting your query.

Please schedule free consultation with Qlik and refer to below link:

https://www.qlik.com/us/products/qlik-sense/ai

From the link. in the bottom use the chat option and book the consultation.

Meanwhile you may also refer to https://www.qlik.com/us/products/qlik-automl

Thanks,

Padma Priya

Senior Technical Support Engineer

Qlik Support

[Albert_Candelario · Accepted solution]

Hello @Likith ,

Here's a high-level overview of the steps to deploy a machine learning model using MLFlow and Azure Machine Learning:

1. Install and set up MLFlow: You can install MLFlow using pip install mlflow. After installation, set up an MLFlow tracking server to keep track of your experiments.

2. Prepare your data and train the model: Split your data into training and testing sets, and then train your model using the training data. Log the experiment information and model parameters using MLFlow.

3. Register the model: Register the trained model with MLFlow, so you can use it in your deployment pipeline.

4. Create an Azure Machine Learning Workspace: If you don't have an existing Azure Machine Learning Workspace, create one and configure it for deployment.

5. Deploy the model to Azure Machine Learning: Use the MLFlow command-line interface (CLI) to deploy the registered model to Azure Machine Learning as a web service.

6. Test the deployed model: Test the deployed model by sending test data to the web service and receiving predictions from it.

7. Monitor and manage the deployed model: Use the Azure Machine Learning Workspace to monitor the deployed model and update it as needed.

This is a general outline of the process, and the actual implementation may vary depending on the specific requirements of your project.

I hope this helps.

Cheers,

Albert

[arthurf]

Does your excellent response refer -- specfically or generally -- to the ability to manage a Qlik AutoML model via MLFlow (in this instance on Azure Machine Learning)? I was planning to create a new thread about if/how to manage Qlik AutoML models in an external instance of MLFlow. Thank you in advance for any assistance or pointing me in the direction to assistance.