课程: MLOps with Databricks

免费学习该课程！

今天就开通帐号，24,600 门业界名师课程任您挑！

Model logging in MLflow

Model logging in MLflow - Databricks教程

课程: MLOps with Databricks

开始一个月试用 购买团队方案

Model logging in MLflow

“

- [Instructor] MLflow supports a lot of different model flavors. Here are some popular ones. Sklearn, pytorch, xgboost, lightgbm and statsmodels. The following functions are supported when working with model flavors. MLflow.<model_flavor>.log_model to log a model is in MLflow artifact for the current run. MLflow.<model_flavor>.load_model to load the model from a local file or an MLflow run, MLflow.<model_flavor>.save_model to save a model to a path in the local file system. Let's build a model using the train set restored as a delta table into get a look. Let's load the train set on the test set as a pySpark data frame. We use an American categorical features in the project config file to predict whether the booking was canceled. Extreme X test, wide train and white test upon the data frames. We have two categorical features, meal type and room type. First of all, we need to transform them using math and coding, for column meal type with values meal plan one, meal plan two, meal plan…

内容