rocky/OSGKeyboard
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
eac44daaf4038695f6ccec2450b68671d288906a
OSGKeyboard/ModelTraining/ClipboardSemantics/corpus-registry-sources.json
T
Rocky aa37067f79 chore(semantics): add v6 release gate pipeline 
- Add reproducible v6 boundary, blessing, and consensus-adjudication
  corpora, plus the tiny-transformer trainer and v6 release-gate
  evaluator that gate every candidate on the deployed baselines.
- Wire consensus-label merging, product-policy anchor evaluation, and
  sealed blessing benchmark review with their pytest coverage.
- Refresh open-training corpus generation, iterative retraining runner,
  and random-holdout evaluation so v6 candidates can be benchmarked
  end-to-end.
2026-08-29 11:51:42 +08:00

303 lines
11 KiB
JSON
Raw Blame History

{
"schemaVersion": 2,
"policy": "Only license-safe sources may produce train candidates. Any text seen in calibration or evaluation data is globally barred from training.",
"intentLabels": [
"task",
"question",
"invitation",
"complaint",
"scheduleNegotiation",
"confirmationDecision",
"followUpReminder",
"blessing",
"replyableMessage",
"assistantCommand",
"informationQuery",
"systemNotification"
],
"domains": [
"finance",
"travel",
"calendar",
"communication",
"media",
"smartHome",
"shopping",
"dining",
"health",
"weather",
"accountService",
"generalKnowledge"
],
"legacyFullyKnownIntentLabels": [
"task",
"question",
"invitation",
"complaint",
"scheduleNegotiation",
"confirmationDecision",
"followUpReminder",
"blessing",
"replyableMessage"
],
"sources": [
{
"id": "historical-nine-model-v1",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/Generated/historical-nine-model-v1.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "by-split",
"sourceType": "project-generated",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "current-product-corpus",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/clipboard_semantic_corpus.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "by-split",
"sourceType": "project-generated",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "product-corpus-16x",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/Generated/product-corpus-16x.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "by-split",
"sourceType": "project-generated",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "licensed-open-training",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/open-training-corpus.jsonl",
"license": "mixed per-record commercial licenses",
"defaultUse": "train",
"sourceType": "licensed-open-data",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": true
},
{
"id": "v6-boundary-synthetic",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/v6-boundary-training-supplement.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "train",
"sourceType": "project-generated-synthetic",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": true
},
{
"id": "v6-migration-tier-a",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/CorpusRegistry/Labels/V6Migration/tier-a.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "train",
"sourceType": "model-consensus",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": true
},
{
"id": "v6-migration-tier-b",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/CorpusRegistry/Labels/V6Migration/tier-b.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "train",
"sourceType": "model-consensus",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": true
},
{
"id": "v6-migration-adjudicated",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/CorpusRegistry/Labels/V6Migration/adjudicated.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "train",
"sourceType": "model-adjudicated",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": true
},
{
"id": "v6-blind-evaluation",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/v6-blind-evaluation-corpus.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "by-split",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": true
},
{
"id": "blessing-synthetic-100k",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/Generated/blessing-synthetic-100k.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "train",
"sourceType": "project-generated",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 0.35,
"required": true
},
{
"id": "consensus-silver-v1",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/Consensus/consensus-silver.jsonl",
"license": "Apache-2.0",
"defaultUse": "by-split",
"sourceType": "model-consensus",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": false
},
{
"id": "random-holdout",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/random-holdout-corpus.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "fresh-metric-holdout",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/fresh-metric-holdout-corpus.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "targeted-confirmation-holdout",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/targeted-confirmation-holdout-corpus.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "targeted-final-holdout",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/targeted-final-holdout-corpus.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "targeted-release-holdout",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/targeted-release-holdout-corpus.jsonl",
"license": "OSGKeyboard project license",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "online-real-holdout",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/online-real-holdout-corpus.jsonl",
"license": "evaluation-only mixed provenance",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "comprehensive-online-holdout",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/comprehensive-online-holdout-corpus.jsonl",
"license": "evaluation-only mixed provenance",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": true,
"knownIntentLabels": ["task", "question", "invitation", "complaint", "scheduleNegotiation", "confirmationDecision", "followUpReminder", "blessing", "replyableMessage"],
"sentimentKnown": true,
"weight": 1.0,
"required": true
},
{
"id": "blessing-benchmark-review",
"path": "${REPO_ROOT}/ModelTraining/ClipboardSemantics/BlessingBenchmark/review-queue.jsonl",
"license": "evaluation-only mixed provenance",
"defaultUse": "evaluation-only",
"sourceType": "blind-review-benchmark",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"sensitive": true,
"required": true
},
{
"id": "blessing-unseen-scenarios",
"path": "${REPO_ROOT}/OSGKeyboardTests/Fixtures/BlessingUnseenScenarios.json",
"license": "OSGKeyboard project license",
"defaultUse": "evaluation-only",
"sourceType": "frozen-holdout",
"allIntentLabelsKnown": false,
"sentimentKnown": false,
"weight": 1.0,
"required": false
}
],
"excludedSources": [
{
"id": "LCCC",
"reason": "Research-only terms and incomplete content-rights/privacy chain. It must not influence production corpus generation or training."
},
{
"id": "DailyDialog-EmpatheticDialogues-Switchboard",
"reason": "Non-commercial license restrictions."
},
{
"id": "CPED",
"reason": "Repository license does not establish commercial rights to the underlying television dialogue."
},
{
"id": "Tianji-Wishes-Birthday-Quotes",
"reason": "No sufficiently clear per-record source or generation-rights chain."
},
{
"id": "CFPB",
"reason": "No official train split; privacy-sensitive narratives overlap the frozen evaluation provenance."
},
{
"id": "CLINC150",
"reason": "Isolated from product training because its assistant taxonomy is not product-policy compatible."
}
]
}
Reference in New Issue View Git Blame Copy Permalink