110 kB

Ctrl+K

4 contributors

History: 15 commits

Codeseys

feat(trainer): policy-optimization objective MENU (ADR-014)

aae66fa about 1 month ago

ADR-001-gpu-venue.md

4.27 kB
Wave 7: Phase 2-4 of deep work loop — backlog, parallel research, three ADRs about 1 month ago
ADR-002-trace-source.md

7.53 kB
Wave 11: cross-model adversarial review + honest down-revision about 1 month ago
ADR-003-diloco-impl.md

5.05 kB
Wave 7: Phase 2-4 of deep work loop — backlog, parallel research, three ADRs about 1 month ago
ADR-004-replaysim-normalization.md

5.76 kB
Wave 13: serverless DiLoCo + replaysim normalization + 3 distillation losses + PRIME-RL + Monarch about 1 month ago
ADR-005-serverless-diloco.md

6.24 kB
Wave 13: serverless DiLoCo + replaysim normalization + 3 distillation losses + PRIME-RL + Monarch about 1 month ago
ADR-006-rl-frameworks.md

6.17 kB
docs(adr): add ADR-008/009/010 (Dr.GRPO+SDPO, layered hints, FeatureDeletionEnv) about 1 month ago
ADR-007-self-distillation-losses.md

12.4 kB
Wave 15: 4-angle multi-model self-critique caught 2 math BLOCKERs in primary loss kernels; fixed against upstream byte-for-byte + GSM8K example + ergonomics about 1 month ago
ADR-008-drgrpo-sdpo-live-channel.md

12.2 kB
review+fix: cross-family adversarial ADR review (owed item) + remediation about 1 month ago
ADR-009-layered-hint-generator.md

9.98 kB
review+fix: cross-family adversarial ADR review (owed item) + remediation about 1 month ago
ADR-010-feature-deletion-datagen.md

14.7 kB
feat(b4-gpu+b6): GPU train-proof on A10G + docker-gated substrate E2E test about 1 month ago
ADR-011-sdpo-alignment-indices.md

5.24 kB
architect: ADR-011/012/013 + research (alignment-index fix, review-findings closure, LMA channel-ladder) about 1 month ago
ADR-012-close-review-findings.md

5.87 kB
architect: ADR-011/012/013 + research (alignment-index fix, review-findings closure, LMA channel-ladder) about 1 month ago
ADR-013-lma-integration-channel-ladder.md

6.66 kB
docs(adr-013): close acceptance-gate boxes 1-5; only user-gated spend remains about 1 month ago
ADR-014-policy-optimization-objective-menu.md

6.09 kB
feat(trainer): policy-optimization objective MENU (ADR-014) about 1 month ago
README.md

1.74 kB
architect: ADR-011/012/013 + research (alignment-index fix, review-findings closure, LMA channel-ladder) about 1 month ago