4a48e99a48
Adds CPU-side read-only GradBalancerMonitor for the existing
`grad_balance_isv_update` GPU kernel. Monitor reads ISV slots 31-35
(GRAD_NORM_TARGET_{DIR,MAG,ORD,URG}_INDEX + GRAD_SCALE_LIMIT_INDEX);
returns mean-of-4-targets as the representative scalar; exposes all 5
slots + mean + fire-rate via diagnose().
Creates the `trainers/dqn/monitors/` module directory. Subsequent tasks
add kernels + monitors for atoms, gamma, kelly_cap, tau, epsilon.
Tests: 3 unit tests pass (read returns mean, diagnose exposes 7 fields,
observe tracks fire-rate with 1e-6 epsilon).
No behavioural change — kernel unchanged, monitor is pure observer.
Plan 1 Task 17. Spec §4.C.6 (2026-04-24 revision).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
ml
10-model ML ensemble for the Foxhunt HFT system, built on Candle v0.9.1.
Models
- DQN (Rainbow) — deep Q-network with prioritized replay, dueling heads, noisy nets
- PPO — proximal policy optimization with GAE, LSTM policies, clip-higher
- TFT — temporal fusion transformer for multi-horizon forecasting
- Mamba2 — state space model for sequence prediction
- Liquid Networks — biologically inspired networks for non-stationary data
- TLOB — transformer-based limit order book analysis
- KAN — Kolmogorov-Arnold networks
- xLSTM — extended LSTM architecture
- TGGN — temporal graph neural network
- Diffusion — diffusion-based generative model
Key Modules
ensemble— model ensemble coordination and confidence aggregationhyperopt— PSO-based hyperparameter optimization with per-model adapterstrainers— unified training loops (DQN, PPO, supervised)inference—InferenceAdaptertrait for predictioncheckpoint— model checkpointing and restorationevaluation— walk-forward evaluation pipeline
Usage
use ml::dqn::DQN;
use ml::ppo::PpoTrainer;