Codeseys
/

composer-replication-framework

Reinforcement Learning

Model card Files Files and versions

composer-replication-framework / composer_replication /trainer /composer_trainer.py

Commit History

Wave 20: Tier-0 fidelity fixes — k1-in-reward KL + Composer-2 behavior rewards

41289bf

Baladithya Balamurugan Claude Opus 4.8 (1M context) commited on 17 days ago

Wave 3: close the HIGH review findings (kill-switch wiring, HeldoutSplit, EKS entrypoint bug)

bd0c358

Baladithya Balamurugan Claude Opus 4.8 (1M context) commited on 17 days ago

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

aae66fa

Codeseys commited on 27 days ago

fix(phase-8): close all 5 cross-family final-verify findings + regression tests

678d10b

Codeseys commited on 28 days ago

feat(wave-a): close ADR-011 (SDPO alignment indices) + ADR-012 (review findings)

d02d724

Codeseys commited on 28 days ago

review+fix: cross-family adversarial ADR review (owed item) + remediation

185cce2

Codeseys commited on 28 days ago

feat(trainer): ADR-008 Dr.GRPO config + SDPO strict-alignment guard

bde5c5e

Codeseys commited on 28 days ago

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

e5add15

Codeseys commited on May 26

Wave 10 — packaging: composer_replication is now pip-installable

ac05fbf

Codeseys commited on May 26