Model Lifecycle¶

Commands for managing models beyond initial training: updating them with new data without full retraining, and governing the promotion pipeline from candidate to production.

/turing:update: Incremental model update¶

Add new data to an existing model without starting from scratch. Model-specific strategies: continued boosting with additional rounds for XGBoost/LightGBM, fine-tuning with reduced learning rate and replay buffer for neural networks, and partial_fit() or warm_start=True for scikit-learn. Includes automatic catastrophic forgetting detection; if performance on old data degrades beyond tolerance, the update is flagged.

Syntax: /turing:update <exp-id> --new-data <path> [--replay-ratio 0.1] [--tolerance 0.005] [--json]

Examples:

/turing:update exp-089 --new-data data/new_batch.csv
/turing:update exp-089 --new-data data/new.csv --replay-ratio 0.2
/turing:update exp-089 --new-data data/new.csv --tolerance 0.01
/turing:update exp-089 --new-data data/new.csv --json

/turing:registry: Model registry¶

Track which model is production, staging, candidate, or archived. Promotion between stages requires passing gates: candidate to staging requires a regression check and seed study; staging to production requires a methodology audit and calibration check. Use --force to skip gate checks when necessary. The registry maintains a full promotion/demotion history for each model.

Syntax: /turing:registry [list|register|promote|demote|archive|history] [exp-id] [stage]

Examples:

/turing:registry list
/turing:registry register exp-095 --version v4.1
/turing:registry promote exp-089 staging
/turing:registry promote exp-089 production --force
/turing:registry demote exp-078 staging --reason "latency regression"
/turing:registry archive exp-042 --reason "superseded by v4"
/turing:registry history
/turing:registry history exp-089