643b9505a4
Final 58 compile errors fixed + GPU violation analysis: - 23 files across ml, ml-dqn, ml-supervised, services - flash_attention: CudaBlas + stream fields, GPU matmul/transpose - ensemble adapters (tggn, tlob, mamba2, tft, ppo): StreamTensor/GpuLinear - diffusion/xlstm/mamba trainable: StreamTensor ↔ GpuTensor conversion - hyperopt adapters: fixed API signatures - trainers (liquid, mamba2): checkpoint save via safetensors - benchmarks: fixed to_scalar, RainbowAgent API - services: MlDevice, PPO::load_checkpoint, Mamba2SSM constructor Host-side softmax/distributional functions analyzed — operating on legitimately-downloaded small output tensors at computation endpoints. Not GPU violations (cold paths, <50 elements). FULL WORKSPACE: 0 errors, 56 warnings, 0 candle dependencies. Co-Authored-By: Claude Opus 4.6 (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;