For that, you’ll need the object of type mlflow.ActiveRun for the current run. In addition, if you wish to load the model soon, it may be convenient to output the run’s ID directly to the console. Import os from random import random, randint from mlflow import log_metric, log_param, log_params, log_artifacts if _name_ = "_main_" : # Log a parameter (key-value pair) log_param ( "config_value", randint ( 0, 100 )) # Log a dictionary of parameters log_params (.log_model.
This example demonstrates the use of these functions:
Mlflow.log_artifacts, mlflow.log_image, mlflow.log_text Values updated during the run (for instance, accuracy)įiles produced by the run (for instance, model weights) In addition, or if you are using a library for which autolog is not yet supported, you may use key-value pairs to track:Ĭonstant values (for instance, configuration parameters) fit ( X_train, y_train ) # Use the model to make predictions on the test dataset. rf = RandomForestRegressor ( n_estimators = 100, max_depth = 6, max_features = 3 ) rf. autolog () db = load_diabetes () X_train, X_test, y_train, y_test = train_test_split ( db. The environment still contains unwanted packages from the pypi channel.Import mlflow from sklearn.model_selection import train_test_split from sklearn.datasets import load_diabetes from sklearn.ensemble import RandomForestRegressor mlflow. Sphinxcontrib-serializinghtml 1.1.4 pypi_0 pypi Sphinxcontrib-applehelp 1.0.2 pypi_0 pypi # packages in environment at /opt/miniconda3:Ĭonda-package-handling 1.6.0 p圓7h1de35cc_0 $ conda list -n base -revisionsĭespite the above conda install command, my base environment is still polluted by various python packages, that were installed after revision 1. $ conda install -n base -revision 1įor reference, here is what revision 1 means in my case. I attempted cleaning up my base (mini)conda environment by installing revision 1 of the base environment.
0 kommentar(er)
