Python Library

Overview of our client library.

Use our python library to instrument your machine learning model and track experiments. Setup should only take a few lines of code. If you're using a popular framework, we have a number of integrations to make setting up wandb easy.

We have more detailed docs generated from the code in Reference.

Instrumenting a model

  • wandb.init — initialize a new run at the top of your training script

  • wandb.config — track hyperparameters

  • wandb.log — log metrics over time within your training loop

  • — save files in association with your run, like model weights

  • wandb.restore — restore the state of your code when you ran a given run