jgrusewski 57bae2cb68 fix(ml): OOM hardening + battle-test KAN/xLSTM/Diffusion models ...

Replace 8 unbounded Vec accumulation patterns with bounded VecDeque
across ensemble, PPO, DQN, Mamba2, and data pipeline code to prevent
OOM on RTX 3050 Ti (4GB VRAM) during live trading and extended training.

Key OOM fixes:
- Ensemble price/volatility history: Vec → VecDeque with O(1) eviction
- Data pipeline: MAX_FEATURES=500K cap (~512MB) prevents unbounded loading
- DQN replay buffer: full-array shuffle → HashSet random sampling (8MB → 256B)
- PPO loss histories: bounded VecDeque (cap 1K), eliminated batch.clone()
- Mamba2 scan: pre-allocated Vecs, explicit drop() after Tensor::cat
- Mamba2 training history: capped at 100, Tensor::randn replaces Vec→Tensor
- Mamba2 SSM reset: 2 unwrap() violations replaced with proper error handling

Battle-testing (19 new integration tests):
- KAN: 5 tests (forward, 50-epoch training 89.9% loss reduction, checkpoint)
- xLSTM: 7 tests (2D+3D forward, 30-epoch training 82% reduction, checkpoint)
- Diffusion: 7 tests (2D+3D forward, 20-epoch pipeline, checkpoint, validation)

Bonus: fix pre-existing cache test failure (match .dbn.zst files, graceful skip)

All 2390 lib tests pass, 0 new clippy errors.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-02-24 09:55:39 +01:00

..

2026-02-20-codebase-deep-clean-design.md

docs: Add codebase deep clean design document

2026-02-20 13:08:10 +01:00

2026-02-20-codebase-deep-clean-implementation.md

docs: Add codebase deep clean implementation plan

2026-02-20 13:21:25 +01:00

2026-02-20-dqn-algorithm-fix-design.md

docs: Add DQN algorithm fix & 2026 modernization design

2026-02-20 14:39:58 +01:00

2026-02-20-dqn-algorithm-fix-implementation.md

docs: Add DQN algorithm fix implementation plan (11 tasks)

2026-02-20 14:44:44 +01:00

2026-02-20-dqn-production-readiness-design.md

docs: add DQN pipeline production readiness design

2026-02-20 20:49:41 +01:00

2026-02-20-dqn-production-readiness-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-20-dqn-training-smoke-test-implementation.md

docs: add DQN training smoke test implementation plan

2026-02-20 20:03:10 +01:00

2026-02-20-dqn-training-smoke-test.md

docs: add DQN training smoke test design

2026-02-20 19:56:20 +01:00

2026-02-20-gradient-accumulation-design.md

docs: add gradient accumulation design

2026-02-20 21:53:45 +01:00

2026-02-20-gradient-accumulation-implementation.md

docs: add gradient accumulation implementation plan

2026-02-20 21:59:14 +01:00

2026-02-20-validation-stack-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-20-validation-stack-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-21-backtesting-vertical-slice-design.md

docs: backtesting vertical slice production readiness design

2026-02-21 15:21:22 +01:00

2026-02-21-backtesting-vertical-slice-implementation.md

docs: backtesting vertical slice implementation plan

2026-02-21 15:30:00 +01:00

2026-02-21-capital-ready-validation-roadmap-design.md

docs: capital-ready validation roadmap design

2026-02-21 10:33:54 +01:00

2026-02-21-capital-ready-validation-roadmap-implementation.md

docs: capital-ready validation roadmap implementation plan

2026-02-21 10:39:54 +01:00

2026-02-21-ensemble-real-inference-design.md

docs: ensemble real inference design and implementation plan

2026-02-21 14:58:36 +01:00

2026-02-21-ensemble-real-inference-implementation.md

docs: ensemble real inference design and implementation plan

2026-02-21 14:58:36 +01:00

2026-02-21-full-stack-ml-integration-design.md

docs: full-stack ML integration design (cleanup → ensemble → hyperopt → paper trading)

2026-02-21 13:19:17 +01:00

2026-02-21-full-stack-ml-integration-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-21-full-system-production-readiness-design.md

docs: full system production readiness design (6-phase plan)

2026-02-21 15:18:19 +01:00

2026-02-21-phase1-data-pipeline-implementation.md

docs: Phase 1-2 data pipeline and trading service implementation plan

2026-02-21 15:25:17 +01:00

2026-02-21-ppo-production-modernization-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-21-ppo-validation-adapter-design.md

feat(validation): add PPO adapter for validation harness (MLP + LSTM)

2026-02-21 10:23:48 +01:00

2026-02-21-production-readiness-phase3-design.md

docs: production readiness phase 3 design — 18 tasks across 4 pillars

2026-02-21 18:53:38 +01:00

2026-02-21-production-readiness-phase3-implementation.md

docs: production readiness phase 3 implementation plan — 18 tasks

2026-02-21 19:00:43 +01:00

2026-02-21-production-readiness-phase4-design.md

docs: Phase 4 production readiness design and implementation plan

2026-02-21 20:21:01 +01:00

2026-02-21-production-readiness-phase4-implementation.md

docs: Phase 4 production readiness design and implementation plan

2026-02-21 20:21:01 +01:00

2026-02-21-production-readiness-phase5-design.md

docs: Phase 5 production readiness design and implementation plan

2026-02-21 23:06:50 +01:00

2026-02-21-production-readiness-phase5-implementation.md

docs: Phase 5 production readiness design and implementation plan

2026-02-21 23:06:50 +01:00

2026-02-21-production-training-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-21-production-training-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-21-real-data-backtest-and-deploy-design.md

docs: real-data ensemble backtest and hyperopt deploy design

2026-02-21 15:04:58 +01:00

2026-02-21-web-dashboard-hybrid-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-baseline-phase3-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-codebase-deduplication-design.md

docs: add codebase de-duplication design

2026-02-22 19:54:53 +01:00

2026-02-22-codebase-deduplication-implementation.md

docs: add codebase de-duplication implementation plan

2026-02-22 20:01:12 +01:00

2026-02-22-dev-git-server-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-dev-git-server-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-documentation-cleanup-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-documentation-cleanup-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-liquid-cfc-modernization-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-liquid-cfc-modernization-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-ml-ensemble-expansion-design.md

docs: add ML ensemble expansion design (7→10 models, 5 integrations)

2026-02-22 23:56:38 +01:00

2026-02-22-production-baseline-implementation.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-production-baseline-integration-tests-design.md

chore: update .gitignore and add design plan docs

2026-02-24 00:30:15 +01:00

2026-02-22-production-hardening-design.md

docs: add production hardening design (53 tasks, 5 phases)

2026-02-22 22:33:08 +01:00

2026-02-22-production-hardening-implementation-phases-3-5.md

docs: add production hardening implementation plan (53 tasks, 5 phases)

2026-02-22 22:51:56 +01:00

2026-02-22-production-hardening-implementation.md

docs: add production hardening implementation plan (53 tasks, 5 phases)

2026-02-22 22:51:56 +01:00

2026-02-22-production-hardening-phase1.md

docs: add production hardening implementation plan (53 tasks, 5 phases)

2026-02-22 22:51:56 +01:00

2026-02-23-data-pipeline-asset-selection-design.md

docs: add data pipeline and asset selection design

2026-02-23 12:28:02 +01:00

2026-02-23-data-pipeline-asset-selection-implementation.md

docs: add data pipeline & asset selection implementation plan

2026-02-23 12:36:45 +01:00

2026-02-23-dynamic-gpu-detection-design.md

docs: add dynamic GPU detection design plan

2026-02-23 10:53:04 +01:00

2026-02-23-dynamic-gpu-detection-implementation.md

docs: add dynamic GPU detection implementation plan

2026-02-23 10:56:45 +01:00

2026-02-23-ml-ensemble-expansion-implementation.md

docs: add ML ensemble expansion implementation plan (36 tasks, 6 phases)

2026-02-23 00:05:28 +01:00

2026-02-23-operational-maturity-design.md

docs: add operational maturity system design (3 pillars)

2026-02-23 12:02:52 +01:00

2026-02-23-operational-maturity-implementation.md

docs: add operational maturity implementation plan (12 tasks, 6 phases)

2026-02-23 12:11:28 +01:00

2026-02-23-production-hardening-phase2-design.md

docs: production hardening phase 2 design — $100K live trading readiness

2026-02-23 09:33:55 +01:00

2026-02-23-production-hardening-phase2-implementation.md

docs: production hardening phase 2 implementation plan — 19 tasks

2026-02-23 09:39:13 +01:00

2026-02-23-production-pipeline-wiring-design.md

docs: add production pipeline wiring design (6 tasks)

2026-02-23 14:43:22 +01:00

2026-02-23-production-pipeline-wiring-implementation.md

docs: add production pipeline wiring implementation plan (10 tasks)

2026-02-23 14:59:18 +01:00

2026-02-23-production-safety-audit-design.md

docs: production safety audit design — 38 fixes across 5 layers

2026-02-23 18:51:54 +01:00

2026-02-23-production-safety-audit-implementation.md

docs: production safety audit implementation plan (38 tasks, 5 layers)

2026-02-23 19:00:29 +01:00

2026-02-23-real-data-training-pipeline-design.md

docs: add real data training pipeline design and implementation plan

2026-02-23 18:07:03 +01:00

2026-02-23-real-data-training-pipeline-implementation.md

docs: add real data training pipeline design and implementation plan

2026-02-23 18:07:03 +01:00

2026-02-23-trading-universe-data-organization-design.md

docs: add trading universe & data organization design

2026-02-23 14:17:22 +01:00

2026-02-23-trading-universe-data-organization-implementation.md

docs: add trading universe & data organization implementation plan

2026-02-23 14:21:48 +01:00

2026-02-24-ml-production-hardening-design.md

fix(ml): OOM hardening + battle-test KAN/xLSTM/Diffusion models

2026-02-24 09:55:39 +01:00