jgrusewski b4cf279b73 fix: reduce H100 replay buffer VRAM fraction 55% → 45% — OOM from IQN 8.6GB + Phase 2 ...

IQN with 64 quantiles × batch=16384 allocates 8.6GB for tiled intermediates.
Combined with attention (250MB), Phase 2 aux workspaces (64MB), backward
branch scratch (16MB), the 55% replay fraction left only 349MB free —
not enough for experience collector's 192MB next_states allocation.

45% frees ~8GB headroom: replay buffer 19.4M → ~15.9M (still 60× batch).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-18 18:42:06 +02:00

..

alternative

chore: Second cleanup wave - organize root directory

2025-10-30 01:26:02 +01:00

base

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

database

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

gpu

fix: reduce H100 replay buffer VRAM fraction 55% → 45% — OOM from IQN 8.6GB + Phase 2

2026-04-18 18:42:06 +02:00

ml

feat: TOML training profile system — config-driven hyperparameters

2026-03-21 23:35:41 +01:00

monitoring

chore(infra): remove dead Cockpit TF, CI-CD dashboards, stale Grafana configs

2026-03-02 15:13:40 +01:00

prometheus

refactor: update K8s, CI, Docker, Prometheus, scripts, and FXT CLI for api rename

2026-03-04 23:46:36 +01:00

redis

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

security

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

sqlx

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

training

feat: cost-driven hold timing — replace min_hold_bars with learned cost signals

2026-04-18 01:04:11 +02:00

tuning

chore: Major documentation cleanup - remove 2,060 obsolete files

2025-11-28 10:29:45 +01:00

development.toml

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

ml_models_staging.toml

feat(wave-d): Complete Phase 6 agents G15-G19 - memory optimization + performance validation

2025-10-18 18:14:34 +02:00

paper_trading_config.yaml

🚀 Wave 160 Phase 5: Complete ML Ensemble + Production Deployment (27 Agents)

2025-10-14 18:41:48 +02:00

performance-benchmark.toml

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

production.toml

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

server-extensions.cnf

chore: Second cleanup wave - organize root directory

2025-10-30 01:26:02 +01:00

staging.toml

🚀 PRODUCTION READY: Foxhunt HFT Trading System v1.0

2025-09-24 23:47:21 +02:00

tarpaulin.toml

chore: Major documentation cleanup - remove 2,060 obsolete files

2025-11-28 10:29:45 +01:00

universe-es-mbp10-test.toml

feat(ml): re-enable OFI features from MBP-10 order book data

2026-03-07 13:00:07 +01:00

universe-es-mbp10.toml

feat(data): schema-configurable Databento download with MinIO upload

2026-03-07 12:19:46 +01:00

universe-es-trades.toml

feat: parameterized Argo WorkflowTemplate for Databento downloads

2026-03-31 22:40:37 +02:00

universe-futures-baseline.toml

fix(ml): add stype_in=Parent for Databento .FUT symbol downloads

2026-02-23 20:02:38 +01:00