mlops-weights-and-biases — Claude Code skill: what it does

What mlops-weights-and-biases does

The skill describes how to use Weights & Biases during model training. It covers recording run settings and metrics, reviewing training progress, comparing experiments, and searching parameter combinations with sweeps. It also covers uploading models and other artifacts so their versions and lineage can be managed.

You need the wandb library and W&B authentication, set up with wandb login or the WANDB_API_KEY environment variable. W&B calls shown in the skill handle logging and checkpoint uploads; your project still supplies the training loop. Without an API key, you cannot authenticate to W&B.