tcapelle/wandb-driven-dev

wandb-driven-dev

Enforce a systematic, reproducible approach to empirical questions — hypothesis first, baseline + variant, smoke before full, wandb as source of truth. Use whenever the user asks 'does X work?', 'is A better than B?', 'what's the best N?', or says…

لغة النص الأصلي: الإنجليزية

wbagent

Use this skill for querying and analyzing Weights & Biases projects through the W&B SDK and the local `wandb_helpers.py` query helpers. Covers run discovery, filtered run-table queries, selected summary metrics, config comparisons, exact run counts,…

لغة النص الأصلي: الإنجليزية