ai-research-reproduction

Purpose

Guide README-first deep learning reproduction toward the smallest trustworthy run with auditable evidence. Preserve documented meaning; record assumptions, deviations and blockers instead of changing semantics to manufacture success. Load specialized references only for a concrete uncertainty.

Fast Path

For a routine bounded run, keep the control path short:

1. Read the target README and only the target test/config/source needed to understand the documented command.
2. Run scripts/orchestrate_repro.py --repo <repo> --plan-only --agent-output with any explicit user timeout bound (--timeout or --train-timeout) already supplied; review command_candidates, the selected cmd-XX, side-effect contract, selection fingerprint, and returned reviewed_run_args. With no --output-dir, later evidence goes to <repo>/repro_outputs regardless of caller cwd.
3. Run the selected candidate, or another reviewed candidate, with --run-selected --command-id <cmd-XX> --plan-fingerprint <fingerprint> --agent-output plus requested timeout/metric/source-adjacent options. Preserve an explicit user command-timeout bound instead of silently making it stricter on a routine trusted run. --timeout limits the target command; do not wrap the whole orchestrator in an equal or shorter external timeout, because it still needs time to terminate children and write terminal evidence. A changed command set fails closed; setup/download commands are never target candidates.
4. Run --verify-output --agent-output; inspect detailed evidence files only when verification fails or the result is partial/blocked.
5. Deliver the bounded result and stop.