weights-and-biases

Call wandb.init with your project name and config, then use wandb.log inside your training loop to record metrics like loss and accuracy. Call wandb.finish at the end, and optionally save checkpoints with wandb.save or as artifacts.

Define a sweep config with a search method (grid, random, or bayes), a target metric, and parameter distributions, then create it with wandb.sweep. Launch agents with wandb.agent pointing to a training function that reads hyperparameters from wandb.config.

Yes, set report_to="wandb" in TrainingArguments and the Trainer automatically logs metrics, hyperparameters, and checkpoints to W&B. You can also add custom WandbCallback subclasses for additional logging.

Yes, set the WANDB_MODE environment variable to "offline" before initializing your run. Metrics are stored locally and can be uploaded later using the wandb sync command on the run directory.

Artifacts are versioned files such as datasets or model checkpoints with automatic lineage tracking between runs. The model registry is a curated collection where model artifacts are linked and promoted through stages like staging and production using aliases.

Bayesian optimization is recommended for expensive training runs because it learns from previous trials to sample promising regions. Use grid search for few discrete parameters and random search for quick exploration across many parameters.