weights-and-biases

Call wandb.init with your project name and config, then call wandb.log inside your training loop to record metrics like loss and accuracy per epoch or batch. Use wandb.watch to automatically log gradients and model parameters, and wandb.finish when training completes.

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 your training function, which reads hyperparameters from wandb.config each trial.

Yes, set report_to="wandb" in TrainingArguments and the HuggingFace Trainer automatically logs metrics, losses, and evaluation results to W&B. You can also add custom WandbCallback subclasses to log additional metrics during evaluation.

Artifacts version any file-based object such as datasets, checkpoints, or predictions with automatic lineage tracking between runs. The model registry is a curated layer where model artifacts are linked and promoted through aliases like staging and production for deployment governance.

Yes, set the WANDB_MODE environment variable to "offline" before calling wandb.init, and all run data is stored locally. Later, run wandb sync on the run directory to upload the logged metrics and artifacts to the W&B server.

Bayesian optimization learns from previous trial results to sample promising hyperparameter regions, making it more sample-efficient for expensive training runs. Grid search exhaustively tests all combinations, which grows exponentially and wastes compute on poor configurations.