MLOps: A Complete Introduction To Monitoring In 26 Mins скачать в хорошем качестве

MLOps: A Complete Introduction To Monitoring In 26 Mins

In this video we cover why we need to monitor ML system and how we can use Evidently to detect drift to prevent model decay. Please follow me on LinkedIn:   / mlopsomari   Link to full code: https://github.com/mlopsomari/evidently 00:00 Intro 00:28 What is machine learning? 03:10 Model decay 04:00 Data drift 04:48 Concept drift 05:55 Why do we need to monitor ML systems? 06:20 Monitoring with Evidently 09:00 Evidently Datasets theory 09:40 Code: Creating Evidently Datasets 11:50 Evidently Evaluations theory 12:50 Code: Creating Evaluations with column-level metrics 15:30 Code: Creating Evaluations with dataset-level metrics 17:00 Code: Creating Evaluations with tests 22:42 Code: Monitoring ML systems with Presets