Sync from GitHub (tests passed)

deep_learning/config.py

@@ -102,7 +102,9 @@ class ASROConfig:
     lambda_quantile: float = 0.4   # w_quantile; was 0.3 (unnormalised old formula)
     # lambda_vol is a sub-weight within the calibration bundle only.
     # It controls how much the Q90-Q10 spread tracks 2× actual σ.
-    lambda_vol: float = 0.2
+    # Two independent Optuna runs (20 trials each) both converged on 0.35 —
+    # updating default to match confirmed optimal value.
+    lambda_vol: float = 0.35
     risk_free_rate: float = 0.0
     sharpe_window: int = 20
 

deep_learning/training/hyperopt.py

@@ -50,7 +50,10 @@ def create_trial_config(trial, base_cfg: TFTASROConfig) -> TFTASROConfig:
         dropout=trial.suggest_float("dropout", 0.1, 0.5, step=0.05),
         hidden_continuous_size=trial.suggest_int("hidden_continuous_size", 8, 32, step=8),
         quantiles=base_cfg.model.quantiles,
-        learning_rate=trial.suggest_float("learning_rate", 5e-5, 5e-3, log=True),
+        # Cap at 1e-3: two consecutive Optuna runs both selected ~3-4e-3 which
+        # caused the model to converge in 1 epoch then diverge. 1e-3 is the
+        # practical upper bound for stable TFT training on ~300 samples.
+        learning_rate=trial.suggest_float("learning_rate", 5e-5, 1e-3, log=True),
         reduce_on_plateau_patience=4,
         gradient_clip_val=trial.suggest_float("gradient_clip_val", 0.5, 2.0, step=0.5),
     )