Replace per-bar GPU forward passes with chunked batch inference (1024 bars
per chunk). This reduces ~90K individual CUDA kernel launches to ~88 batched
forward passes — ~1000× fewer GPU round-trips.
Changes:
- Add DQN::batch_greedy_actions(&self) for immutable batched forward+argmax
- Add RegimeConditionalDQN::batch_greedy_actions with per-regime-head batching
- Add DQNTrainer::convert_to_state_vec (CPU-only, skips GPU tensor allocation)
- Add PortfolioTracker::set_position_direct for backtest state sync
- Rewrite hyperopt backtest loop: chunked batching with portfolio state
updates between chunks (1024-bar granularity ≈ 17h of 1-min data)
Portfolio features (last 3 of 54 dims) are refreshed between chunks via
set_portfolio_for_backtest(), keeping position/PnL/exposure accurate at
chunk boundaries while batching inference within each chunk.
2640 tests pass, 0 clippy warnings.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Split model overhead into two constants: MODEL_OVERHEAD_MB (pure
model weights for batch-size capping) and TRIAL_VRAM_MB (total
per-trial VRAM for concurrent hyperopt planning). DQN trials
empirically consume ~7 GB each on L40S (model + GPU replay buffer +
experience collector + CUDA allocations + fragmentation), not the
200 MB previously estimated. This caused plan_hyperopt to compute
128 concurrent trials instead of the actual 5, inflating PSO
particles from 20→128 and total trials from 20→384 via .max()
instead of .min(), guaranteeing a 4h timeout kill.
Fix auto-scaling to: (1) match particles to GPU concurrency for
maximum hardware utilization on any node, (2) cap particles at
max_trials to never inflate the trial budget, (3) never auto-inflate
the total trial count.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The GPU experience collection path bypassed monitor.track_action(),
leaving action_counts all zeros and reporting 0/45 diversity on every
epoch. Feed batch.actions into monitor.action_counts after GPU
collection so the metric reflects actual policy behavior.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The fxt client was using a hardcoded dev fallback secret for JWT signing,
causing InvalidSignature errors against the API gateway. Now reads
jwt_secret from ~/.foxhunt/config.toml with fallback chain:
env var > config file > dev secret.
Also updates default api_gateway_url to https://api.fxhnt.ai.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Swap the stale trading_service monitoring proto (10 RPCs, never
implemented) for the real monitoring_service proto (2 RPCs:
GetLiveTrainingMetrics, StreamTrainingMetrics). Add a dedicated
MonitoringServiceProxy that forwards directly to monitoring-service
on port 50057, so `fxt watch` works end-to-end through the gateway.
- build.rs: compile monitoring_service/proto/monitoring.proto with server+client
- MonitoringServiceProxy: new zero-copy proxy (unary + streaming)
- TradingServiceProxy: remove monitoring_client, stub 6 stale methods
- main.rs: MONITORING_SERVICE_URL env, optional proxy with graceful degradation
- Network policies: api-gateway↔monitoring-service egress/ingress on 50057
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The Monte Carlo entropy estimate with random (untrained) weights can dip
well below -1.0 under parallel test load. Widened the lower bound from
-1.0 to -5.0; the key invariant is finiteness, not positivity.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Rename SlippageModel trait → MarketImpactModel to avoid name collision
with PPO's SlippageModel enum
- Wrap ThompsonSamplingState in std::sync::Mutex for interior mutability
through Arc, enabling record_outcome during inference
- Add TrafficSplitter::record_outcome() method for callers
- Cap RegimeDetectionEngine::feature_data to window_size*3 entries
- Add trailing newline to ab_testing.rs
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Extends TrafficSplitter with Thompson Sampling (Beta-distributed traffic
allocation) and PromotionCriteria for automated champion/challenger
decisions. 18 tests, 0 clippy.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Per-trade online learning with Elastic Weight Consolidation to prevent
catastrophic forgetting. Rolling 10K experience buffer, mini-updates
every 100 trades. Safety rails: grad clip 1.0, LR×0.1, auto-rollback
at 20% Sharpe degradation, kill switch at Sharpe < -1.0.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
SlippageModel trait with two implementations:
- LinearImpactModel: sqrt market impact (spread/2 + eta*sqrt(V/ADV) + order premium)
- FixedCostModel: backward-compat wrapper for existing 15/5/10 bps constants
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add PositionFeatures struct that extracts 3 features for RL agent
position awareness (54→57 dim state vector):
- unrealized_pnl: normalized by EMA volatility
- bars_in_position: log-scaled to [0, 1]
- cost_basis: relative to price in bps, clamped [-500, 500]
8 unit tests, 0 clippy warnings. Not yet wired into trainers.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
DQN: wire SAC-style entropy regularizer into TD loss, fix diversity
penalty normalization from ln(3) to ln(45), add epsilon floor (0.05)
for noisy nets, enable sigma scheduling (0.8→0.4 with 0.3 floor),
add UCB count-based exploration bonus, expand hyperopt search space
from 43D to 45D.
PPO: replace fabricated entropy metric (value_loss×0.5) with real
Shannon entropy from action distribution, fix EntropyRegularizer
normalization from ln(3) to ln(45), implement Monte Carlo entropy
estimate for flow policy (was returning zeros), fix hyperopt VRAM
bound from 3 to 45 actions.
Monitoring: add normalized action entropy and diversity Prometheus
gauges, add exploration diagnostics logging per epoch.
2503 ml tests pass, 277 common tests pass, 0 clippy warnings.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- ml_training_service: handlers delegate to promotion_manager
- api_gateway: proxy pass-through for both new RPCs
- Proto messages added to service-side ml_training.proto
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
DashboardState with 4 tab substates (Training/Trading/Risk/System),
ring buffers for loss history and fills, scroll support.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- ratatui 0.29, crossterm 0.28 (event-stream), tokio-stream to deps
- ApproveModel + RejectModel RPCs added to MLTrainingService
- 4 new proto messages for model promotion workflow
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Ratatui streaming dashboard (fxt watch) with 4 tabs, stub command
wiring plan, and new ApproveModel/RejectModel proto RPCs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
5 tests: training loop cycle, priority-weighted sampling, KS distribution
test, IS weight range validation, ring buffer overwrite correctness.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Collect loss tensors on GPU during accumulation loop and perform a
single mean+to_scalar readback at the accumulation boundary instead
of N per-step CPU readbacks. Falls back to CPU sum when GPU tensors
are unavailable. Also adds missing loss_tensor_gpu field to
RegimeConditionalDQN's GradientResult construction.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
GradientResult and ComputeLossResult gain td_errors_gpu/indices_gpu
Option<Tensor> fields for GPU-resident TD errors. Trainer accumulates
GPU tensors and calls update_priorities_gpu() when available, falling
back to CPU path otherwise. DqnAgent delegates via memory().
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
New variant delegates to GpuReplayBuffer for all operations. Sample
returns BatchSample with gpu_batch populated (empty CPU vecs). Added
update_priorities_gpu() for tensor-based priority updates, as_gpu_buffer()
for direct access from trainer. Existing tests unchanged.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Proportional: cumsum + binary search over alpha-weighted priorities.
Rank-based: sort descending, rank probabilities 1/rank^alpha, cumsum sample.
Priority update: |td_error|^alpha + epsilon scatter into buffer.
All methods return GpuBatch with IS weights normalized by max.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Slice-scatter based insertion handles wrap-around by splitting into
tail + head copies. New experiences get max_priority. Tests verify
correctness at capacity boundary.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Pre-allocates all experience tensors on device at init (~47 MB for 100K
buffer). Ring buffer cursor and beta annealing state tracked CPU-side.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
GPU-resident batch tensors allow compute_gradients() to skip CPU→GPU
transfer when sampling from the GPU replay buffer.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
PerformanceMetrics::from_trades() already returns win_rate as percentage
(e.g. 57.78), but TRIAL_SUMMARY and backtest details log lines multiplied
by 100 again, producing values like 5778%. Remove the extra * 100.0.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
TDD plan covering GpuReplayBuffer, proportional + rank-based GPU
sampling, priority scatter updates, async loss readback, trainer
integration, OOM fallback, and distribution correctness tests.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
