<!-- agent-source-envelope -->
# Artifact: Factory Run Schema v1 | PTLLM

> The canonical target, data, post-training, evaluation, package, and report shape for PostTrainLLM runs.

Canonical: https://posttrainllm.com/artifacts/factory-run-schema-v1
<!-- /agent-source-envelope -->

Process artifact

Report artifact

2026-07-02

A specialist factory needs proof folders, not vibes. This schema defines the local run directory, required metrics, and decision vocabulary that every public artifact should eventually satisfy.

factory, evals, reports

## Headline Numbers

Evidence class applies to each value: measured = retained evaluation or
runtime result; historical = recorded result not freshly reproduced;
derived = calculated or decided from records; observed = current artifact
state; not-measured = an explicit evidence gap.

**Required files:** 8

config, dataset, train log, evals, report, artifact, decision

Evidence: derived

**Decisions:** 6

ship, reject, retry-data, retry-training, retry-eval, park

Evidence: derived

**First-class outputs:** 5

data, training, eval, package, report

Evidence: derived

## Competitive Context

| System | Metric | Score | Size / Class | Comparable? | Readout |
| --- | --- | --- | --- | --- | --- |
| posttrainllm factory schema | required public run files | 8 | repo-local contract | Direct | Defines the minimum evidence bundle each future public model artifact must carry. |
| Ad hoc model card only | before/after reproducibility | weak | single document | Directional | Useful for release notes, but insufficient for a factory claim without eval JSON, decision, and blockers. |

Direct rows share this artifact's eval setup. Directional rows are
useful market context but should not be read as leaderboard claims.

## Run folder contract

| File | Purpose | Public relevance |
| --- | --- | --- |
| config.json | Target, base, method, thresholds | Explains what was attempted |
| dataset.json | Sources, rows, filtering, heldout | Provenance |
| eval-baseline.json | Frozen baseline result | Before number |
| eval-candidate.json | Candidate result | After number |
| decision.json | Ship/reject/retry call | Honest release status |

## Release Blockers

### Live commands do not emit the whole schema automatically

The routed SQL report card is now the canonical rendered example, but training and evaluation commands still require orchestration to assemble every run fragment.

**Unblock:** Make future train/eval commands emit schema-complete run folders and enforce them with publish-check.

## Evidence

- [Run schema](https://posttrainllm.com/docs/factory/run-schema)

- [Report template](https://posttrainllm.com/docs/factory/reports)

- [Packaging rules](https://posttrainllm.com/docs/factory/packaging)

- [Canonical routed SQL report card](https://posttrainllm.com/report-cards/qwen06-sql-routed-v1)

## Next Release Action

Use the routed SQL report as the canonical example; make live commands emit the same complete contract automatically.
