gitadmin/foxhunt
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(sp22-vnext): F-3c follow-up — K=3 CE EMA in stdout HEALTH_DIAG aux line

Browse Source 
The F-3c commit (cdd3dc6ed) pushed `aux_trade_outcome_ce_ema` into
the regression-detection metrics vec but never wired a `tracing::
info!` console emit. Smoke `train-q5k5k` ran clean past the
rollout + post-rollout phases — but `argo logs train-q5k5k | grep
aux_trade_outcome` returned zero hits, so the operator had no way
to read the K=3 head's CE EMA trajectory.

Fix: extend the existing K=2 aux HEALTH_DIAG print

  HEALTH_DIAG[N]: aux [next_bar_mse=… regime_ce=… w=…]

to include `trade_outcome_ce=…`:

  HEALTH_DIAG[N]: aux [next_bar_mse=… regime_ce=…
                       trade_outcome_ce=… w=…]

Reads ISV[538] via the same `trainer.read_isv_signal_at(...)`
accessor the regression-detection vec uses — single source of
truth, no duplicated mirroring.

Expected operator-visible trajectory across smoke epochs:
  Epoch 0 (cold-start, Pearl A sentinel):     0.000
  Epoch 1+ (first non-zero bootstrap):       ~1.099 (= ln(3))
  Healthy learning:                           0.5 - 0.7
  Pinned at ln(3) for many epochs:            head can't learn

cargo check -p ml clean. Audit doc updated.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
jgrusewski
2026-05-14 15:47:01 +02:00
parent 256a5fa5ab
commit 878cc9ba72
2 changed files with 51 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
crates/ml/src/trainers/dqn/trainer/training_loop.rs
View File

@@ -5994,20 +5994,32 @@ impl DQNTrainer {
use crate::cuda_pipeline::gpu_dqn_trainer::{
AUX_NEXT_BAR_MSE_EMA_INDEX, AUX_REGIME_CE_EMA_INDEX,
};
let (aux_nb_mse, aux_rg_ce, aux_w) =
use crate::cuda_pipeline::sp22_isv_slots::AUX_TRADE_OUTCOME_CE_EMA_INDEX;
let (aux_nb_mse, aux_rg_ce, aux_to_ce, aux_w) =
if let Some(ref fused) = self.fused_ctx {
let trainer = fused.trainer();
(
trainer.read_isv_signal_at(AUX_NEXT_BAR_MSE_EMA_INDEX),
trainer.read_isv_signal_at(AUX_REGIME_CE_EMA_INDEX),
// SP22 H6 vNext Phase F-3c follow-up (2026-05-14):
// K=3 trade-outcome head CE EMA in the existing aux
// print line. The F-3c commit pushed this metric to
// the regression-detection vec but never added it to
// a `tracing::info!` line, so smoke v3 (train-q5k5k)
// ran without operator-visible CE feedback. Cold-
// start = 0.000 (Pearl A sentinel); after first
// trade-close batch = ~ln(3) ≈ 1.099 (uniform
// K=3 prior); healthy learning trajectory drops
// toward 0.5-0.7 across epochs.
trainer.read_isv_signal_at(AUX_TRADE_OUTCOME_CE_EMA_INDEX),
trainer.aux_weight(),
)
} else {
(0.0, 0.0, 0.0)
(0.0, 0.0, 0.0, 0.0)
};
tracing::info!(
"HEALTH_DIAG[{}]: aux [next_bar_mse={:.3e} regime_ce={:.3e} w={:.3}]",
epoch, aux_nb_mse, aux_rg_ce, aux_w,
"HEALTH_DIAG[{}]: aux [next_bar_mse={:.3e} regime_ce={:.3e} trade_outcome_ce={:.3e} w={:.3}]",
epoch, aux_nb_mse, aux_rg_ce, aux_to_ce, aux_w,
);
}

35
docs/dqn-wire-up-audit.md
View File

@@ -2,6 +2,41 @@
**Status:** Populated during Plan 1 Task 6 (A.5 orphan audit). Updated on every commit per Invariant 7.
## 2026-05-14 — feat(sp22-vnext): F-3c follow-up — K=3 CE EMA in stdout HEALTH_DIAG line
**Branch:** `sp20-aux-h-fixed`. **Trigger:** smoke `train-q5k5k`
ran clean past the rollout + post-rollout phases, but no
`aux_trade_outcome_ce` value visible in any tracing log because
the F-3c commit (`cdd3dc6ed`) only pushed the metric to the
regression-detection vec (`metric_bands.update_and_check`), never
added it to a `tracing::info!` console emit.
The F-3c commit message claimed "Operator sees CE every epoch:
Cold-start: 0.000 / ~1.098 (= ln(3)) / ~0.5-0.7" — but `argo logs
train-q5k5k | grep aux_trade_outcome` returned zero hits because
the metric never reached stdout.
**Fix:** extend the existing K=2 aux HEALTH_DIAG print
(`HEALTH_DIAG[N]: aux [next_bar_mse=… regime_ce=… w=…]`) to
include `trade_outcome_ce=…`:
```rust
tracing::info!(
"HEALTH_DIAG[{}]: aux [next_bar_mse={:.3e} regime_ce={:.3e} \
trade_outcome_ce={:.3e} w={:.3}]",
epoch, aux_nb_mse, aux_rg_ce, aux_to_ce, aux_w,
);
```
Reads ISV[538] via the same `trainer.read_isv_signal_at(...)`
accessor the regression-detection vec uses — single source of
truth.
**Tests:** `cargo check -p ml --lib` clean.
**Touched:** `crates/ml/src/trainers/dqn/trainer/training_loop.rs`
(single print-line edit + new import).
## 2026-05-14 — fix(sp22-vnext): K=3 aux_outcome_labels CF-half buffer underrun
**Branch:** `sp20-aux-h-fixed`. **Trigger:** workflow `train-xzv56`