jgrusewski bfc3ffa9dc diag(sp15-wave5): in-capture CAPTURE_PHASE_* checkpoints ...

Smoke train-fp7xx printed all 16 Phase-6/7/8 checkpoints clean through
PER_PRIORITY_DONE. CAPTURE_DONE missing, exit changed 139→143 (SIGTERM)
— capture_training_graph hangs or takes ~40s past PER_PRIORITY_DONE
before Argo terminates the pod.

Adds 14 CAPTURE_PHASE_* checkpoints across the 12 child captures +
parent compose:
  BEGIN / PER_SAMPLE / COUNTERS / SPECTRAL / FORWARD / DDQN / AUX
  POST_AUX / ADAM_GRAD / ADAM_UPDATE / MAINTENANCE / IQL_MODULATE
  PER_PRIORITY / CHILDREN_STORED / PARENT_COMPOSED

Diagnostic-only.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-07 12:47:23 +02:00

..

asset_selection

…

backtesting

…

benchmark

…

checkpoint

…

cuda_common

…

cuda_pipeline

fix(cuda): compute_expected_q stride 13 vs denoise_target_q_buf size 12 — OOB writes threads 60-63

2026-05-07 11:39:34 +02:00

data_loaders

cleanup: declarative rewrites for deferred-work TODOs in ml crate

2026-04-23 08:41:44 +02:00

data_pipeline

…

data_validation

…

deployment

…

diffusion

…

dqn

…

ensemble

fix(ensemble): DQN adapter tests use correct STATE_DIM input size

2026-04-23 10:50:56 +02:00

evaluation

…

explainability

…

features

fix(data): structural defense at data-loading boundary — 5 layers

2026-05-02 14:02:53 +02:00

flash_attention

…

gpu_benchmarks

…

hyperopt

feat(sp15-wave3b): host-side wire-up — eliminates 5 orphan launchers via GpuBacktestEvaluator constructor signature change

2026-05-06 23:39:45 +02:00

integration

…

kan

…

labeling

…

liquid

…

mamba

…

microstructure

…

model_registry

…

observability

…

paper_trading

…

ppo

…

regime

…

regime_detection

…

registry

…

risk

…

security

…

stress_testing

…

tests

…

tft

…

tgnn

…

tlob

…

trainers

diag(sp15-wave5): in-capture CAPTURE_PHASE_* checkpoints

2026-05-07 12:47:23 +02:00

training

audit(dqn-v2): A.5 orphan audit complete

2026-04-24 13:10:49 +02:00

transformers

cleanup: declarative rewrites for deferred-work TODOs in ml crate

2026-04-23 08:41:44 +02:00

universe

…

validation

…

xlstm

…

.gitignore

…

batch_processing.rs

…

benchmarks.rs

…

bridge.rs

…

data_loader.rs

chore(ml): gate ZN.FUT data-loader tests + migrate test HtoD/DtoH to mapped-pinned

2026-05-03 11:34:41 +02:00

feature_cache.rs

refactor(data_source): default to mbp10 across smoke + localdev profiles

2026-04-27 22:35:47 +02:00

fxcache.rs

fix(data): target stride/column drift — host + GPU consumer cleanup

2026-05-02 19:48:49 +02:00

inference.rs

…

lib.rs

refactor(ml): drop 8 leaf sub-crates from trading + backtesting services

2026-05-01 01:45:18 +02:00

model_factory.rs

…

model_registry.rs

…

preprocessing.rs

…

training_pipeline.rs

…

training_profile.rs

feat(sp4): Layer B — atomic consumer migration to ISV-driven bounds

2026-05-01 11:10:41 +02:00

training.rs

audit(dqn-v2): A.5 orphan audit complete

2026-04-24 13:10:49 +02:00

walk_forward.rs

fix(data): structural defense at data-loading boundary — 5 layers

2026-05-02 14:02:53 +02:00