foxhunt

Files

jgrusewski 55aeddaebd feat(ml-alpha): anchor_l2 kernel + Wiener-α controller (v2 B) [V8]

L2 anchor regularization toward initialization (axis B). Anchors
horizon_tokens + Q + MoE experts toward their init values to prevent
the calibration drift observed in v1 (where val_loss climbed as α
opened past epoch 1 in 2 of 3 folds).

KERNEL (`anchor_l2_fwd_bwd`):
  loss_out = λ · Σ_i (p[i] − p_init[i])²
  grad_p[i] += 2λ · (p[i] − p_init[i])

  - Warp-shuffle reduce; one block per parameter group; strided thread
    loop over n. Cross-warp reduce uses one __syncthreads.
  - Coalesced grad write via stride loop.
  - λ passed as device-side [1]-buffer (host writes scalar before launch
    — capture-safe).

CONTROLLER (`trainer::anchor_controller::AnchorController`):
  - Signal-driven λ floor: λ_floor = ‖p_init‖ / (100 · √numel).
    Cross-fold-persistent per pearl_kelly_cap_signal_driven_floors.
  - Wiener-α smoother (α = diff_var / (diff_var + sample_var + ε))
    on val_loss change; α floored at 0.4 per
    pearl_wiener_alpha_floor_for_nonstationary.
  - λ blends toward target = |ema_change|·scale with α; floored at
    λ_floor per pearl_blend_formulas_must_have_permanent_floor.
  - First-observation bootstrap (sentinel state replaced directly on
    first step) per pearl_first_observation_bootstrap.
  - 4/4 unit tests PASS: signal-floor init, bootstrap returns floor,
    floor protection across 1000 steps, λ_max cap.

NUMGRAD VERIFICATION (RTX 3050 sm_86):
  anchor_l2_numgrad PASSES with closed-form parity (machine precision)
  and central-difference parity (4 random positions) within 5e-2 rel.

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

2026-05-18 14:26:31 +02:00

adamw_step.cu

perf(ml-alpha): device-resident AdamW step counter (capture prep)

2026-05-17 14:39:38 +02:00

anchor_l2.cu

feat(ml-alpha): anchor_l2 kernel + Wiener-α controller (v2 B) [V8]

2026-05-18 14:26:31 +02:00

attention_pool.cu

perf(ml-alpha): block-per-batch attention pool bwd refactor (Phase B commit 4)

2026-05-18 00:01:31 +02:00

bce_loss_multi_horizon_sigma.cu

feat(ml-alpha): Kendall sigma-weighted BCE kernel + numgrad (v2 A) [V7]