foxhunt

Files

jgrusewski 28384836a7 tune: batch 8192→16384, lr 1e-5→2e-5, tau 0.005→0.007

Linear scaling rule: 2x batch → 2x LR to maintain effective update
magnitude. Tau increased to compensate for fewer steps/epoch (target
network tracks faster). H100 VRAM: 41GB free at B=8192, B=16384 adds
~4GB — easily fits.

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

2026-04-19 15:38:31 +02:00

dqn-hyperopt.toml

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

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

dqn-localdev.toml

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

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

dqn-production.toml

tune: batch 8192→16384, lr 1e-5→2e-5, tau 0.005→0.007

2026-04-19 15:38:31 +02:00

dqn-smoketest.toml

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