jgrusewski 8dfd49bda1 fix(rl): move gates + log_pi OUTSIDE CUDA Graph capture ...

The confidence gate and FRD gate were captured as no-ops during graph
warmup (step 1, before warmup threshold). Graph replay then never
fired them. Moving outside the capture region makes them execute as
individual kernel launches every step.

Result: Hold=100% from step 200-1000 (gate enforcing surfer Hold
default). Gate fires on nearly every non-Hold action until Q
confidence exceeds threshold=0.3.

Hold still drops after ~1500 steps as Q becomes confident — need
higher/adaptive threshold for sustained Hold protection.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

2026-05-27 01:36:59 +02:00

..

cuda

feat(rl): confidence gate on ALL non-Hold + post-gate log_pi + surfer fixes

2026-05-27 01:30:01 +02:00

examples

fix(rl): wire BCE + aux losses into GPU loader path + event-based sync

2026-05-26 22:18:30 +02:00

src

fix(rl): move gates + log_pi OUTSIDE CUDA Graph capture

2026-05-27 01:36:59 +02:00

tests

perf(rl): fuse adamw_increment into adamw_step + tau into iqn_forward

2026-05-26 16:46:10 +02:00

build.rs

perf(rl): GPU-resident data loader — zero CPU per step

2026-05-26 20:51:06 +02:00

Cargo.toml

feat(rl): R6 — replace LobEnv with kernel-driven GPU-pure env step

2026-05-23 10:39:25 +02:00