Updated thesis

src/trainers/diffusion_trainer.py

@@ -242,30 +242,31 @@ class DiffusionTrainer:
 
         _, generated_sampels = self.test(test_loader, -1, task)
         # self.policy_evaluator.plot_profits_table()
+        if self.policy_evaluator:
 
-        optimal_penalty, profit, charge_cycles = (
-            self.policy_evaluator.optimize_penalty_for_target_charge_cycles(
-                idx_samples=generated_sampels,
-                test_loader=test_loader,
-                initial_penalty=900,
-                target_charge_cycles=283,
-                initial_learning_rate=1,
-                max_iterations=50,
-                tolerance=1,
+            optimal_penalty, profit, charge_cycles = (
+                self.policy_evaluator.optimize_penalty_for_target_charge_cycles(
+                    idx_samples=generated_sampels,
+                    test_loader=test_loader,
+                    initial_penalty=900,
+                    target_charge_cycles=283,
+                    initial_learning_rate=1,
+                    max_iterations=50,
+                    tolerance=1,
+                )
             )
-        )
 
-        print(
-            f"Optimal Penalty: {optimal_penalty}, Profit: {profit}, Charge Cycles: {charge_cycles}"
-        )
+            print(
+                f"Optimal Penalty: {optimal_penalty}, Profit: {profit}, Charge Cycles: {charge_cycles}"
+            )
 
-        task.get_logger().report_single_value(
-            name="Optimal Penalty", value=optimal_penalty
-        )
-        task.get_logger().report_single_value(name="Optimal Profit", value=profit)
-        task.get_logger().report_single_value(
-            name="Optimal Charge Cycles", value=charge_cycles
-        )
+            task.get_logger().report_single_value(
+                name="Optimal Penalty", value=optimal_penalty
+            )
+            task.get_logger().report_single_value(name="Optimal Profit", value=profit)
+            task.get_logger().report_single_value(
+                name="Optimal Charge Cycles", value=charge_cycles
+            )
 
         if task:
             task.close()
@@ -436,13 +437,13 @@ class DiffusionTrainer:
                     inversed_samples_batched - inversed_expanded_targets
                 )
                 inversed_mae_mean = inversed_mae.mean()
-                all_inversed_mae.extend(inversed_mae_mean.tolist())
+                all_inversed_mae.append(inversed_mae_mean)
 
                 inversed_mse = np.square(
                     inversed_samples_batched - inversed_expanded_targets
                 )
                 inversed_mse_mean = inversed_mse.mean()
-                all_inversed_mse.extend(inversed_mse_mean.tolist())
+                all_inversed_mse.append(inversed_mse_mean)
 
             # add all values from crps_mean to all_crps
             all_crps.extend(crps_mean.tolist())
@@ -460,12 +461,12 @@ class DiffusionTrainer:
 
             mean_inversed_mae = np.array(all_inversed_mae).mean()
             task.get_logger().report_single_value(
-                name="test_MSELoss", value=mean_inversed_mae
+                name="test_L1Loss", value=mean_inversed_mae
             )
 
             mean_inversed_mse = np.array(all_inversed_mse).mean()
             task.get_logger().report_single_value(
-                name="test_L1Loss", value=mean_inversed_mse
+                name="test_MSELoss", value=mean_inversed_mse
             )
 
         if self.best_score is None or mean_crps < self.best_score:

src/training_scripts/diffusion_training.py

@@ -2,7 +2,7 @@ from src.utils.clearml import ClearMLHelper
 
 clearml_helper = ClearMLHelper(project_name="Thesis/NrvForecast")
 task = clearml_helper.get_task(
-    task_name="Diffusion Training: hidden_sizes=[256, 256], lr=0.0001, time_dim=8"
+    task_name="Diffusion Training: hidden_sizes=[256, 256], lr=0.0001, time_dim=8 + Load + PV + Wind + NP"
 )
 task.execute_remotely(queue_name="default", exit_process=True)
 
@@ -19,16 +19,19 @@ from src.policies.PolicyEvaluator import PolicyEvaluator
 data_config = DataConfig()
 data_config.NRV_HISTORY = True
 
-data_config.LOAD_HISTORY = False
-data_config.LOAD_FORECAST = False
+data_config.LOAD_HISTORY = True
+data_config.LOAD_FORECAST = True
 
-data_config.WIND_FORECAST = False
-data_config.WIND_HISTORY = False
+data_config.PV_FORECAST = True
+data_config.PV_HISTORY = True
 
-data_config.QUARTER = False
-data_config.DAY_OF_WEEK = False
+data_config.WIND_FORECAST = True
+data_config.WIND_HISTORY = True
 
-data_config.NOMINAL_NET_POSITION = False
+data_config.QUARTER = True
+data_config.DAY_OF_WEEK = True
+
+data_config.NOMINAL_NET_POSITION = True
 
 data_config = task.connect(data_config, name="data_features")
 
@@ -42,7 +45,7 @@ print("Input dim: ", inputDim)
 model_parameters = {
     "epochs": 15000,
     "learning_rate": 0.0001,
-    "hidden_sizes": [256, 256],
+    "hidden_sizes": [256, 256, 256],
     "time_dim": 8,
 }
 
@@ -70,7 +73,5 @@ baseline_policy = BaselinePolicy(battery, data_path="")
 policy_evaluator = PolicyEvaluator(baseline_policy, task)
 
 #### Trainer ####
-trainer = DiffusionTrainer(
-    model, data_processor, "cuda", policy_evaluator=policy_evaluator
-)
+trainer = DiffusionTrainer(model, data_processor, "cuda", policy_evaluator=None)
 trainer.train(model_parameters["epochs"], model_parameters["learning_rate"], task)