gitadmin/foxhunt
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2,535 Commits 13 Branches 0 Tags
af940671bc488ddc2d90c4e5867a14165e797e86
Commit Graph

6 Commits

Author SHA1 Message Date
jgrusewski
af940671bc feat: reward config pipeline — DQNHyperparameters + TOML profiles 
Add 7 composite reward fields to DQNHyperparameters: w_dsr, w_pnl,
w_dd, w_idle, dd_threshold, loss_aversion, time_decay_rate.

Add RewardSection to training_profile.rs with Option<f64> fields and
apply_to() mapping. Add [reward] section to all 3 DQN TOML profiles
(production, smoketest, hyperopt) with identical defaults.

Remove hold_reward from ExperienceSection (replaced by w_idle).
Add 7 reward search bounds to SearchSpaceSection and bound() match.
Add 7 reward phase_fast defaults to PhaseFastSection.

hold_penalty kept as deprecated field for hyperopt adapter compat
(Task 4 will clean it up).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-03-22 20:50:23 +01:00
jgrusewski
6219491db6 refactor: move smoke test config to dqn-smoketest.toml, restore check_err 
Smoke test loads all hyperparams from TOML profile instead of hardcoding.
TOML: hidden_dim=64, batch=64, lr=0.0003 (stable on RTX 3050 + H100).

Restored check_err() drain in device.rs — required to clear stale CUDA
errors from primary context reuse between tests.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-03-22 12:07:49 +01:00
jgrusewski
43998a330a feat: two-phase hyperopt + backtest evaluator VRAM leak fix 
Two-phase hyperopt splits 31D PSO search into sequential phases:
- Phase 1 (--phase fast, default): fix architecture to small network
  (hidden_dim=128, num_atoms=11), search learning dynamics (~15D).
- Phase 2 (--phase full): fix dynamics from Phase 1 JSON, search
  architecture (~5D). Halves dimensionality per phase → better convergence.
- Phase 1 output includes best_continuous_vector for Phase 2 consumption.

GpuBacktestEvaluator Drop impl: sync forked stream, destroy CUDA graph
and cuBLAS handles before CudaSlice buffers drop. Fixes 261MB/trial
VRAM leak on H100 hyperopt.

ml-core clippy fixes: hex literal, remove dead check_err drain,
unnecessary safety comment, unused OnceLock import.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-03-22 10:32:37 +01:00
jgrusewski
e97c30b50a perf: cap hyperopt hidden_dim=256, num_atoms=51 — 4x faster trials 
Production defaults are sufficient for hyperopt exploration. Larger networks
can be tested in a separate phase with the best hyperparams found.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-03-22 10:00:26 +01:00
jgrusewski
3c8e177932 feat: HyperoptProfile with TOML search space bounds 
Add SearchSpaceSection, PsoSection, HyperoptProfile structs to
training_profile.rs. All 31 PSO search bounds now configurable in
config/training/dqn-hyperopt.toml — no code changes needed to
adjust search ranges.

HyperoptProfile::bound("field", default) returns the TOML value
or falls back to the hardcoded default. Adapter wiring is next step.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-03-22 09:33:55 +01:00
jgrusewski
a75a98bd0d feat: TOML training profile system — config-driven hyperparameters 
Training Profile Loader:
- 3-tier resolution: $FOXHUNT_TRAINING_PROFILE > filesystem > embedded defaults
- DqnTrainingProfile with 10 sections, all Option<T> for sparse profiles
- apply_to() applies only Some fields, preserving struct defaults
- 11 unit tests, all passing

TOML Profiles (config/training/):
- dqn-production.toml: full Rainbow DQN (40+ params)
- dqn-smoketest.toml: CI fast path (sparse, 8 overrides)
- dqn-hyperopt.toml: PSO search space ranges + fixed flags
- ppo-production.toml, ppo-smoketest.toml
- supervised-production.toml, supervised-smoketest.toml
- walk-forward.toml: window sizes

CLI Integration:
- train_baseline_rl: --training-profile (default: dqn-production)
- train_baseline_supervised: --training-profile (default: supervised-production)
- Merge priority: CLI args > TOML profile > GPU profile > struct defaults

Smoke Tests:
- smoke_params() now loads dqn-smoketest.toml instead of hardcoding
- Production features set as manual overrides (testing flags, not config)

Infrastructure:
- K8s job-template.yaml: TRAINING_PROFILE env var + --training-profile arg
- Delete old config/ml/training.toml (replaced, zero callers)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-03-21 23:35:41 +01:00