The W&B client is an open source library and CLI (wandb) for organizing and analyzing your machine learning experiments. Think of it as a framework-agnostic lightweight TensorBoard that persists additional information such as the state of your code, system metrics, and configuration parameters.

• Store config parameters used in a training run
• Associate version control with your training runs
• Search, compare, and visualize training runs
• Analyze system usage metrics alongside runs
• Collaborate with team members
• Run parameter sweeps
• Persist runs forever

pip install wandb

In your training script:

import wandb
# Your custom arguments defined here
args = ...

run = wandb.init(config=args)
run.config["more"] = "custom"

def training_loop():
    while True:
        # Do some machine learning
        epoch, loss, val_loss = ...
        # Framework agnostic / custom metrics
        wandb.log({"epoch": epoch, "loss": loss, "val_loss": val_loss})

Run wandb signup from the directory of your training script. If you already have an account, you can run wandb init to initialize a new directory. You can checkin wandb/settings to version control to share your project with other users.

Run your script with python my_script.py and all metadata will be synced to the cloud. Data is staged locally in a directory named wandb relative to your script. If you want to test your script without syncing to the cloud you can run wandb off.

Framework specific and detailed usage can be found in our documentation.