Plotted controller max normalzied across epoch. Also training average normalized

analysis/controller_across_epochs.py

@@ -9,6 +9,13 @@ from multiprocessing import Pool, cpu_count
 # Define PendulumController class
 from PendulumController import PendulumController
 
+# Constants
+g = 9.81  # Gravity
+R = 1.0   # Length of the pendulum
+m = 10.0  # Mass
+dt = 0.02  # Time step
+num_steps = 500  # Simulation time steps
+
 # ODE solver (RK4 method)
 def pendulum_ode_step(state, dt, desired_theta, controller):
     theta, omega, alpha = state
@@ -44,41 +51,9 @@ def pendulum_ode_step(state, dt, desired_theta, controller):
     new_state = state + (k1 + 2*k2 + 2*k3 + k4) / 6.0
     return new_state
 
-# Constants
+def run_simulation(params):
-g = 9.81  # Gravity
+    controller_file, initial_condition = params
-R = 1.0   # Length of the pendulum
+    theta0, omega0, alpha0, desired_theta = initial_condition
-m = 10.0   # Mass
-dt = 0.02  # Time step
-num_steps = 500  # Simulation time steps
-
-# Directory containing controller files
-loss_function = "quadratic"
-controller_dir = f"/home/judson/Neural-Networks-in-GNC/inverted_pendulum/training/normalized/training/{loss_function}/controllers"
-#controller_dir = f"C:/Users/Judson/Desktop/New Gitea/Neural-Networks-in-GNC/inverted_pendulum/training/{loss_function}/controllers"
-controller_files = sorted([f for f in os.listdir(controller_dir) if f.startswith("controller_") and f.endswith(".pth")])
-
-# Sorting controllers by epoch
-controller_epochs = [int(f.split('_')[1].split('.')[0]) for f in controller_files]
-sorted_controllers = [x for _, x in sorted(zip(controller_epochs, controller_files))]
-
-# **Epoch Range Selection**
-epoch_range = (0, 1000)  # Set your desired range (e.g., (0, 5000) or (0, 100))
-
-filtered_controllers = [
-    f for f in sorted_controllers
-    if epoch_range[0] <= int(f.split('_')[1].split('.')[0]) <= epoch_range[1]
-]
-
-# **Granularity Control: Select every Nth controller**
-N = 1  # Change this value to adjust granularity (e.g., every 5th controller)
-selected_controllers = filtered_controllers[::N]  # Take every Nth controller within the range
-
-# Initial condition
-# theta0, omega0, alpha0, desired_theta = (-np.pi, -2*np.pi, 0.0, -1.3*np.pi)  # Example initial condition
-theta0, omega0, alpha0, desired_theta = (-np.pi, 0.0, 0.0, 0.0)  # Example initial condition
-
-# Parallel function must return epoch explicitly
-def run_simulation(controller_file):
     epoch = int(controller_file.split('_')[1].split('.')[0])
     
     # Load controller
@@ -96,44 +71,62 @@ def run_simulation(controller_file):
 
     return epoch, theta_vals  # Return epoch with data
 
-# Parallel processing
+# Named initial conditions
+initial_conditions = {
+    "small_perturbation": (0.1*np.pi, 0.0, 0.0, 0.0),
+    "large_perturbation": (-np.pi, 0.0, 0.0, 0),
+    "overshoot_vertical_test": (-0.1*np.pi, 2*np.pi, 0.0, 0.0),
+    "overshoot_angle_test": (0.2*np.pi, 2*np.pi, 0.0, 0.3*np.pi),
+    "extreme_perturbation": (4*np.pi, 0.0, 0.0, 0),
+}
+
+# Loss functions to iterate over
+loss_functions = ["constant", "linear", "quadratic", "exponential", "inverse", "inverse_squared"]
+
+
+epoch_start = 0   # Start of the epoch range
+epoch_end = 500  # End of the epoch range
+epoch_step = 5    # Interval between epochs
+
 if __name__ == "__main__":
+    for condition_name, initial_condition in initial_conditions.items():
+        full_path = f"/home/judson/Neural-Networks-in-GNC/inverted_pendulum/analysis/max_normalized/{condition_name}"
+        os.makedirs(full_path, exist_ok=True)  # Create directory if it does not exist
+        
+        for loss_function in loss_functions:
+            controller_dir = f"/home/judson/Neural-Networks-in-GNC/inverted_pendulum/training/normalized/max_normalized/{loss_function}/controllers"
+            controller_files = sorted([f for f in os.listdir(controller_dir) if f.startswith("controller_") and f.endswith(".pth")])
+            # Extract epoch numbers and filter based on the defined range and interval
+            epoch_numbers = [int(f.split('_')[1].split('.')[0]) for f in controller_files]
+            selected_epochs = [e for e in epoch_numbers if epoch_start <= e <= epoch_end and (e - epoch_start) % epoch_step == 0]
+
+            # Filter the controller files to include only those within the selected epochs
+            selected_controllers = [f for f in controller_files if int(f.split('_')[1].split('.')[0]) in selected_epochs]
+            selected_controllers.sort(key=lambda f: int(f.split('_')[1].split('.')[0]))
+
+            # Setup parallel processing
             num_workers = min(cpu_count(), 16)  # Limit to 16 workers max
-    print(f"Using {num_workers} parallel workers...")
+            print(f"Using {num_workers} parallel workers for {loss_function} with initial condition {condition_name}...")
 
             with Pool(processes=num_workers) as pool:
-        results = pool.map(run_simulation, selected_controllers)
+                params = [(controller_file, initial_condition) for controller_file in selected_controllers]
+                results = pool.map(run_simulation, params)
 
-    # Sort results by epoch to ensure correct order
             results.sort(key=lambda x: x[0])
-    epochs, theta_over_epochs = zip(*results)  # Unzip sorted results
+            epochs, theta_over_epochs = zip(*results)
 
-    # Convert results to NumPy arrays
+            fig = plt.figure(figsize=(7, 5))
-    theta_over_epochs = np.array(theta_over_epochs)
-
-    # Create 3D line plot
-    fig = plt.figure(figsize=(10, 7))
             ax = fig.add_subplot(111, projection='3d')
+            time_steps = np.arange(num_steps) * dt
 
-    time_steps = np.arange(num_steps) * dt  # X-axis (time)
+            # Plot the epochs in reverse order because we view it where epoch 0 is in front
+            for epoch, theta_vals in reversed(list(zip(epochs, theta_over_epochs))):
+                ax.plot([epoch] * len(time_steps), time_steps, theta_vals)
 
-    # Plot each controller as a separate line
-    for epoch, theta_vals in zip(epochs, theta_over_epochs):
-        ax.plot(
-            [epoch] * len(time_steps),  # Y-axis (epoch stays constant for each line)
-            time_steps,  # X-axis (time)
-            theta_vals,  # Z-axis (theta evolution)
-            label=f"Epoch {epoch}" if epoch % (N * 10) == 0 else "",  # Label some lines for clarity
-        )
-
-    # Labels
-    ax.set_xlabel("Epoch")
-    ax.set_ylabel("Time (s)")
-    ax.set_zlabel("Theta (rad)")
-    ax.set_title(f"Pendulum Angle Evolution for {loss_function}")
 
             # Add a horizontal line at desired_theta across all epochs and time steps
             epochs_array = np.array([epoch for epoch, _ in zip(epochs, theta_over_epochs)])
+            desired_theta = initial_condition[-1]
             ax.plot(
                 epochs_array,  # X-axis spanning all epochs
                 [time_steps.max()] * len(epochs_array),  # Y-axis at the maximum time step
@@ -141,9 +134,37 @@ if __name__ == "__main__":
                 color='r', linestyle='--', linewidth=2, label='Desired Theta at End Time'
             )
 
-    # Improve visibility
+            ax.set_xlabel("Epoch")
+            ax.set_ylabel("Time (s)")
+            ax.set_zlabel("Theta (rad)")
+            condition_text = f"IC_{'_'.join(map(lambda x: str(round(x, 2)), initial_condition))}"
+            ax.set_title(f"Pendulum Angle Evolution for {loss_function} and {condition_text}")
+
+            # Calculate the range of theta values across all epochs
+            theta_values = np.concatenate(theta_over_epochs)
+            theta_min = np.min(theta_values)
+            theta_max = np.max(theta_values)
+
+            # Determine the desired range around the desired_theta
+            desired_range_min = desired_theta - 1 * np.pi
+            desired_range_max = desired_theta + 1 * np.pi
+
+            # Check if current theta values fall outside the desired range
+            if theta_min < desired_range_min:
+                desired_range_min = desired_range_min
+            else:
+                desired_range_min = theta_min
+
+            if theta_max > desired_range_max:
+                desired_range_max = desired_range_max
+            else:
+                desired_range_max = theta_max
+            
+            ax.set_zlim(desired_range_min, desired_range_max)
+
             ax.view_init(elev=20, azim=-135)  # Adjust 3D perspective
 
-    plt.savefig(f"{loss_function}.png", dpi=600)
+            plot_filename = os.path.join(full_path, f"{loss_function}.png")
-    #plt.show()
+            plt.savefig(plot_filename, dpi=300)
-    print(f"Saved plot as '{loss_function}.png'.")
+            plt.close()
+            print(f"Saved plot as '{plot_filename}'.")

analysis/controller_across_epochs_old.py

@@ -0,0 +1,149 @@
+import os
+import numpy as np
+import torch
+import torch.nn as nn
+import matplotlib.pyplot as plt
+from mpl_toolkits.mplot3d import Axes3D
+from multiprocessing import Pool, cpu_count
+
+# Define PendulumController class
+from PendulumController import PendulumController
+
+# ODE solver (RK4 method)
+def pendulum_ode_step(state, dt, desired_theta, controller):
+    theta, omega, alpha = state
+
+    def compute_torque(th, om, al):
+        inp = torch.tensor([[th, om, al, desired_theta]], dtype=torch.float32)
+        with torch.no_grad():
+            torque = controller(inp)
+            torque = torch.clamp(torque, -250, 250)
+        return float(torque)
+
+    def derivatives(state, torque):
+        th, om, al = state
+        a = (g / R) * np.sin(th) + torque / (m * R**2)
+        return np.array([om, a, 0])  # dtheta, domega, dalpha
+
+    # Compute RK4 steps
+    torque1 = compute_torque(theta, omega, alpha)
+    k1 = dt * derivatives(state, torque1)
+
+    state_k2 = state + 0.5 * k1
+    torque2 = compute_torque(state_k2[0], state_k2[1], state_k2[2])
+    k2 = dt * derivatives(state_k2, torque2)
+
+    state_k3 = state + 0.5 * k2
+    torque3 = compute_torque(state_k3[0], state_k3[1], state_k3[2])
+    k3 = dt * derivatives(state_k3, torque3)
+
+    state_k4 = state + k3
+    torque4 = compute_torque(state_k4[0], state_k4[1], state_k4[2])
+    k4 = dt * derivatives(state_k4, torque4)
+
+    new_state = state + (k1 + 2*k2 + 2*k3 + k4) / 6.0
+    return new_state
+
+# Constants
+g = 9.81  # Gravity
+R = 1.0   # Length of the pendulum
+m = 10.0   # Mass
+dt = 0.02  # Time step
+num_steps = 500  # Simulation time steps
+
+# Directory containing controller files
+loss_function = "quadratic"
+controller_dir = f"/home/judson/Neural-Networks-in-GNC/inverted_pendulum/training/normalized/training/{loss_function}/controllers"
+#controller_dir = f"C:/Users/Judson/Desktop/New Gitea/Neural-Networks-in-GNC/inverted_pendulum/training/{loss_function}/controllers"
+controller_files = sorted([f for f in os.listdir(controller_dir) if f.startswith("controller_") and f.endswith(".pth")])
+
+# Sorting controllers by epoch
+controller_epochs = [int(f.split('_')[1].split('.')[0]) for f in controller_files]
+sorted_controllers = [x for _, x in sorted(zip(controller_epochs, controller_files))]
+
+# **Epoch Range Selection**
+epoch_range = (0, 1000)  # Set your desired range (e.g., (0, 5000) or (0, 100))
+
+filtered_controllers = [
+    f for f in sorted_controllers
+    if epoch_range[0] <= int(f.split('_')[1].split('.')[0]) <= epoch_range[1]
+]
+
+# **Granularity Control: Select every Nth controller**
+N = 1  # Change this value to adjust granularity (e.g., every 5th controller)
+selected_controllers = filtered_controllers[::N]  # Take every Nth controller within the range
+
+# Initial condition
+# theta0, omega0, alpha0, desired_theta = (-np.pi, -2*np.pi, 0.0, -1.3*np.pi)  # Example initial condition
+theta0, omega0, alpha0, desired_theta = (-np.pi, 0.0, 0.0, 0.0)  # Example initial condition
+
+# Parallel function must return epoch explicitly
+def run_simulation(controller_file):
+    epoch = int(controller_file.split('_')[1].split('.')[0])
+    
+    # Load controller
+    controller = PendulumController()
+    controller.load_state_dict(torch.load(os.path.join(controller_dir, controller_file)))
+    controller.eval()
+    
+    # Run simulation
+    state = np.array([theta0, omega0, alpha0])
+    theta_vals = []
+
+    for _ in range(num_steps):
+        theta_vals.append(state[0])
+        state = pendulum_ode_step(state, dt, desired_theta, controller)
+
+    return epoch, theta_vals  # Return epoch with data
+
+# Parallel processing
+if __name__ == "__main__":
+    num_workers = min(cpu_count(), 16)  # Limit to 16 workers max
+    print(f"Using {num_workers} parallel workers...")
+    
+    with Pool(processes=num_workers) as pool:
+        results = pool.map(run_simulation, selected_controllers)
+
+    # Sort results by epoch to ensure correct order
+    results.sort(key=lambda x: x[0])  
+    epochs, theta_over_epochs = zip(*results)  # Unzip sorted results
+
+    # Convert results to NumPy arrays
+    theta_over_epochs = np.array(theta_over_epochs)
+
+    # Create 3D line plot
+    fig = plt.figure(figsize=(10, 7))
+    ax = fig.add_subplot(111, projection='3d')
+
+    time_steps = np.arange(num_steps) * dt  # X-axis (time)
+
+    # Plot each controller as a separate line
+    for epoch, theta_vals in zip(epochs, theta_over_epochs):
+        ax.plot(
+            [epoch] * len(time_steps),  # Y-axis (epoch stays constant for each line)
+            time_steps,  # X-axis (time)
+            theta_vals,  # Z-axis (theta evolution)
+            label=f"Epoch {epoch}" if epoch % (N * 10) == 0 else "",  # Label some lines for clarity
+        )
+
+    # Labels
+    ax.set_xlabel("Epoch")
+    ax.set_ylabel("Time (s)")
+    ax.set_zlabel("Theta (rad)")
+    ax.set_title(f"Pendulum Angle Evolution for {loss_function}")
+
+    # Add a horizontal line at desired_theta across all epochs and time steps
+    epochs_array = np.array([epoch for epoch, _ in zip(epochs, theta_over_epochs)])
+    ax.plot(
+        epochs_array,  # X-axis spanning all epochs
+        [time_steps.max()] * len(epochs_array),  # Y-axis at the maximum time step
+        [desired_theta] * len(epochs_array),  # Constant Z-axis value of desired_theta
+        color='r', linestyle='--', linewidth=2, label='Desired Theta at End Time'
+    )
+
+    # Improve visibility
+    ax.view_init(elev=20, azim=-135)  # Adjust 3D perspective
+
+    plt.savefig(f"{loss_function}.png", dpi=600)
+    #plt.show()
+    print(f"Saved plot as '{loss_function}.png'.")

training/normalized/PendulumController.py

@@ -0,0 +1,17 @@
+import torch
+import torch.nn as nn
+
+class PendulumController(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.net = nn.Sequential(
+            nn.Linear(4, 64),
+            nn.ReLU(),
+            nn.Linear(64, 64),
+            nn.ReLU(),
+            nn.Linear(64, 1)
+        )
+
+    def forward(self, x):
+        raw_torque = self.net(x)
+        return torch.clamp(raw_torque, -250, 250)

training/normalized/PendulumDynamics.py

@@ -0,0 +1,26 @@
+import torch
+import torch.nn as nn
+
+class PendulumDynamics(nn.Module):
+    def __init__(self, controller, m:'float'=1, R:'float'=1, g:'float'=9.81):
+        super().__init__()
+        self.controller = controller
+        self.m: 'float' = m
+        self.R: 'float' = R
+        self.g: 'float' = g
+
+    def forward(self, t, state):
+        # Get the current values from the state
+        theta, omega, alpha, desired_theta = state[:, 0], state[:, 1], state[:, 2], state[:, 3]
+
+        # Make the input stack for the controller
+        input = torch.stack([theta, omega, alpha, desired_theta], dim=1)
+
+        # Get the torque (the output of the neural network)
+        tau = self.controller(input).squeeze(-1)
+
+        # Relax alpha
+        alpha_desired = (self.g / self.R) * torch.sin(theta) + tau / (self.m * self.R**2)
+        dalpha = alpha_desired - alpha
+        
+        return torch.stack([omega, alpha, dalpha, torch.zeros_like(desired_theta)], dim=1)

training/normalized/average_normalized/constant/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/average_normalized/exponential/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/average_normalized/inverse/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/average_normalized/inverse_squared/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/average_normalized/inverse_squared/training_log.csv

@@ -0,0 +1,608 @@
+Epoch,Loss
+0,733.2171630859375
+1,747.2186279296875
+2,173.32571411132812
+3,113.02052307128906
+4,52.3023796081543
+5,29.964460372924805
+6,22.173477172851562
+7,17.20752716064453
+8,14.393138885498047
+9,13.495450019836426
+10,12.72439193725586
+11,12.004809379577637
+12,11.283649444580078
+13,10.60710620880127
+14,9.974536895751953
+15,9.39864730834961
+16,8.96867847442627
+17,8.593585968017578
+18,8.431589126586914
+19,8.36359977722168
+20,8.257978439331055
+21,8.131260871887207
+22,7.987368583679199
+23,7.828018665313721
+24,7.654465198516846
+25,7.468804836273193
+26,7.282363414764404
+27,7.1038055419921875
+28,6.926774024963379
+29,6.756581783294678
+30,6.605012893676758
+31,6.480014324188232
+32,6.387184143066406
+33,6.323246955871582
+34,6.280386924743652
+35,6.2463226318359375
+36,6.210468769073486
+37,6.165651798248291
+38,6.10935640335083
+39,6.044148921966553
+40,5.975587844848633
+41,5.909937381744385
+42,5.85064697265625
+43,5.797428131103516
+44,5.749576091766357
+45,5.7068305015563965
+46,5.668132305145264
+47,5.631911754608154
+48,5.5969085693359375
+49,5.562211513519287
+50,5.527288436889648
+51,5.491824626922607
+52,5.455735206604004
+53,5.4193949699401855
+54,5.383573055267334
+55,5.350064754486084
+56,5.319507598876953
+57,5.289746284484863
+58,5.259618759155273
+59,5.228979110717773
+60,5.198094844818115
+61,5.167768955230713
+62,5.1394476890563965
+63,5.113860607147217
+64,5.0905303955078125
+65,5.070118427276611
+66,5.05070686340332
+67,5.031530857086182
+68,5.012385845184326
+69,4.993386268615723
+70,4.974751949310303
+71,4.956557273864746
+72,4.939281940460205
+73,4.923177719116211
+74,4.9080328941345215
+75,4.893461227416992
+76,4.879059791564941
+77,4.86448860168457
+78,4.8493499755859375
+79,4.833917617797852
+80,4.818392753601074
+81,4.80300760269165
+82,4.788031101226807
+83,4.773463249206543
+84,4.759511947631836
+85,4.746511459350586
+86,4.734636306762695
+87,4.72379207611084
+88,4.713471412658691
+89,4.7037153244018555
+90,4.694935321807861
+91,4.687223434448242
+92,4.680534362792969
+93,4.674868106842041
+94,4.670145511627197
+95,4.666247367858887
+96,4.6630682945251465
+97,4.660345554351807
+98,4.6577630043029785
+99,4.655178070068359
+100,4.65253210067749
+101,4.649806499481201
+102,4.64708137512207
+103,4.644402980804443
+104,4.641839504241943
+105,4.639404296875
+106,4.637064456939697
+107,4.634809970855713
+108,4.632608890533447
+109,4.63045597076416
+110,4.6283416748046875
+111,4.626272201538086
+112,4.624259948730469
+113,4.622305870056152
+114,4.620428085327148
+115,4.618651866912842
+116,4.616915702819824
+117,4.615218162536621
+118,4.613545894622803
+119,4.6118950843811035
+120,4.610260963439941
+121,4.608641147613525
+122,4.607039928436279
+123,4.6054606437683105
+124,4.603922367095947
+125,4.602435111999512
+126,4.601007461547852
+127,4.599658966064453
+128,4.598391532897949
+129,4.59716796875
+130,4.595986843109131
+131,4.594844341278076
+132,4.593727111816406
+133,4.592639446258545
+134,4.591578960418701
+135,4.590541839599609
+136,4.589527130126953
+137,4.588528633117676
+138,4.587544918060303
+139,4.586572647094727
+140,4.585610389709473
+141,4.58465576171875
+142,4.583714008331299
+143,4.582784652709961
+144,4.58186149597168
+145,4.580949783325195
+146,4.58004903793335
+147,4.579157829284668
+148,4.578279972076416
+149,4.5774149894714355
+150,4.5765604972839355
+151,4.575720310211182
+152,4.574892044067383
+153,4.574074745178223
+154,4.573271751403809
+155,4.57248067855835
+156,4.571700096130371
+157,4.570932388305664
+158,4.570174694061279
+159,4.569428443908691
+160,4.568695545196533
+161,4.567972183227539
+162,4.567262649536133
+163,4.566561698913574
+164,4.565873622894287
+165,4.565197944641113
+166,4.564536094665527
+167,4.563884258270264
+168,4.563241958618164
+169,4.562610626220703
+170,4.561987400054932
+171,4.561375141143799
+172,4.56077241897583
+173,4.560176849365234
+174,4.559585094451904
+175,4.558999061584473
+176,4.558421611785889
+177,4.557851791381836
+178,4.55728816986084
+179,4.556730270385742
+180,4.556179523468018
+181,4.555635452270508
+182,4.555098533630371
+183,4.554568767547607
+184,4.554043769836426
+185,4.553525447845459
+186,4.553014755249023
+187,4.552509307861328
+188,4.552009105682373
+189,4.551515102386475
+190,4.551027297973633
+191,4.550544738769531
+192,4.55006742477417
+193,4.549595832824707
+194,4.549131393432617
+195,4.5486741065979
+196,4.548222064971924
+197,4.547779083251953
+198,4.547341823577881
+199,4.546910762786865
+200,4.546485424041748
+201,4.546065330505371
+202,4.545650959014893
+203,4.545242786407471
+204,4.544842720031738
+205,4.5444488525390625
+206,4.544060230255127
+207,4.54367733001709
+208,4.543309211730957
+209,4.5429511070251465
+210,4.542597770690918
+211,4.542244911193848
+212,4.541895389556885
+213,4.541547775268555
+214,4.541203498840332
+215,4.540863513946533
+216,4.540529727935791
+217,4.5402045249938965
+218,4.539884090423584
+219,4.539568901062012
+220,4.539254665374756
+221,4.538944721221924
+222,4.538638591766357
+223,4.538336277008057
+224,4.538040637969971
+225,4.537751197814941
+226,4.537464618682861
+227,4.537179470062256
+228,4.536898612976074
+229,4.536619186401367
+230,4.536343574523926
+231,4.536070823669434
+232,4.535802364349365
+233,4.535538196563721
+234,4.535276889801025
+235,4.535017967224121
+236,4.534761428833008
+237,4.534510135650635
+238,4.5342607498168945
+239,4.534013271331787
+240,4.533768653869629
+241,4.533526420593262
+242,4.533287525177002
+243,4.533048629760742
+244,4.532813549041748
+245,4.532580375671387
+246,4.532350063323975
+247,4.5321245193481445
+248,4.531898021697998
+249,4.531674861907959
+250,4.531455039978027
+251,4.531236171722412
+252,4.531019687652588
+253,4.530806064605713
+254,4.530594348907471
+255,4.530384540557861
+256,4.530177593231201
+257,4.529972076416016
+258,4.529767990112305
+259,4.529566287994385
+260,4.529365539550781
+261,4.529166221618652
+262,4.5289692878723145
+263,4.528773784637451
+264,4.5285797119140625
+265,4.52838659286499
+266,4.528195858001709
+267,4.528006553649902
+268,4.52781867980957
+269,4.527633190155029
+270,4.527448654174805
+271,4.527266502380371
+272,4.527083396911621
+273,4.52690315246582
+274,4.526723861694336
+275,4.526546001434326
+276,4.526369571685791
+277,4.526192665100098
+278,4.526019096374512
+279,4.525848388671875
+280,4.525679588317871
+281,4.5255126953125
+282,4.5253472328186035
+283,4.525182723999023
+284,4.525018692016602
+285,4.524855613708496
+286,4.524693489074707
+287,4.524532794952393
+288,4.524373531341553
+289,4.524214267730713
+290,4.5240559577941895
+291,4.523898601531982
+292,4.523742198944092
+293,4.523586273193359
+294,4.523431301116943
+295,4.52327823638916
+296,4.523126602172852
+297,4.522974491119385
+298,4.522823333740234
+299,4.5226731300354
+300,4.522523403167725
+301,4.522375106811523
+302,4.522228240966797
+303,4.522083282470703
+304,4.521938800811768
+305,4.521796703338623
+306,4.521655559539795
+307,4.521514415740967
+308,4.521374225616455
+309,4.521234035491943
+310,4.521094799041748
+311,4.520956993103027
+312,4.520818710327148
+313,4.520682334899902
+314,4.520545959472656
+315,4.520410060882568
+316,4.520275115966797
+317,4.5201416015625
+318,4.520009517669678
+319,4.519878387451172
+320,4.519747734069824
+321,4.519618034362793
+322,4.51948881149292
+323,4.519359588623047
+324,4.519231796264648
+325,4.519104957580566
+326,4.518978118896484
+327,4.518850803375244
+328,4.5187249183654785
+329,4.518599510192871
+330,4.518474578857422
+331,4.518350601196289
+332,4.518225193023682
+333,4.518101215362549
+334,4.517977237701416
+335,4.517852783203125
+336,4.51772928237915
+337,4.517606258392334
+338,4.517481803894043
+339,4.517359733581543
+340,4.517236709594727
+341,4.517114162445068
+342,4.51699161529541
+343,4.516868591308594
+344,4.516744613647461
+345,4.516623020172119
+346,4.516499996185303
+347,4.5163774490356445
+348,4.516254901885986
+349,4.5161333084106445
+350,4.5160112380981445
+351,4.5158891677856445
+352,4.515769004821777
+353,4.5156474113464355
+354,4.51552677154541
+355,4.515407085418701
+356,4.515288829803467
+357,4.515170574188232
+358,4.515053749084473
+359,4.5149383544921875
+360,4.514822483062744
+361,4.514708042144775
+362,4.5145955085754395
+363,4.514482498168945
+364,4.514371395111084
+365,4.5142598152160645
+366,4.5141496658325195
+367,4.514039993286133
+368,4.513929843902588
+369,4.513820171356201
+370,4.513709545135498
+371,4.513599872589111
+372,4.51348876953125
+373,4.5133771896362305
+374,4.513265132904053
+375,4.513152122497559
+376,4.51303768157959
+377,4.512922763824463
+378,4.512807846069336
+379,4.512691974639893
+380,4.512577056884766
+381,4.512462139129639
+382,4.512347221374512
+383,4.512232780456543
+384,4.512118339538574
+385,4.5120038986206055
+386,4.511888027191162
+387,4.511774063110352
+388,4.511661052703857
+389,4.5115485191345215
+390,4.511435508728027
+391,4.511322975158691
+392,4.5112104415893555
+393,4.511097431182861
+394,4.510985851287842
+395,4.510873317718506
+396,4.510761260986328
+397,4.510649681091309
+398,4.510537624359131
+399,4.510426998138428
+400,4.510315895080566
+401,4.510205268859863
+402,4.510096073150635
+403,4.509986400604248
+404,4.5098772048950195
+405,4.509767532348633
+406,4.509658336639404
+407,4.509550094604492
+408,4.509440898895264
+409,4.509332180023193
+410,4.509223937988281
+411,4.509115219116211
+412,4.509006500244141
+413,4.508897304534912
+414,4.508789539337158
+415,4.508681774139404
+416,4.508574485778809
+417,4.508468151092529
+418,4.508361339569092
+419,4.5082550048828125
+420,4.508149147033691
+421,4.5080437660217285
+422,4.507938385009766
+423,4.5078325271606445
+424,4.507728099822998
+425,4.507623672485352
+426,4.507519245147705
+427,4.507416725158691
+428,4.5073137283325195
+429,4.507210731506348
+430,4.50710916519165
+431,4.507007122039795
+432,4.506906032562256
+433,4.506804466247559
+434,4.506704330444336
+435,4.50660514831543
+436,4.506505966186523
+437,4.506407260894775
+438,4.5063090324401855
+439,4.50621223449707
+440,4.506114959716797
+441,4.50601863861084
+442,4.505922794342041
+443,4.5058274269104
+444,4.50573205947876
+445,4.505638122558594
+446,4.505544185638428
+447,4.505451679229736
+448,4.505359172821045
+449,4.505267143249512
+450,4.505176067352295
+451,4.505085468292236
+452,4.504995822906494
+453,4.504907131195068
+454,4.504818439483643
+455,4.504730701446533
+456,4.504644393920898
+457,4.504558086395264
+458,4.504471778869629
+459,4.504387378692627
+460,4.504302024841309
+461,4.5042195320129395
+462,4.504136085510254
+463,4.504052639007568
+464,4.503970623016357
+465,4.503889083862305
+466,4.503807544708252
+467,4.503726005554199
+468,4.503644943237305
+469,4.503564357757568
+470,4.503483295440674
+471,4.503403186798096
+472,4.503323078155518
+473,4.503242015838623
+474,4.5031633377075195
+475,4.5030837059021
+476,4.50300407409668
+477,4.502924919128418
+478,4.502845764160156
+479,4.5027666091918945
+480,4.502687931060791
+481,4.5026092529296875
+482,4.502531051635742
+483,4.502452373504639
+484,4.502374172210693
+485,4.502295970916748
+486,4.502218246459961
+487,4.502140045166016
+488,4.50206184387207
+489,4.501983642578125
+490,4.501905918121338
+491,4.501828193664551
+492,4.5017499923706055
+493,4.501672744750977
+494,4.5015950202941895
+495,4.501518249511719
+496,4.501441955566406
+497,4.501365661621094
+498,4.5012898445129395
+499,4.501214981079102
+500,4.501140117645264
+501,4.501065254211426
+502,4.500990867614746
+503,4.500916004180908
+504,4.5008416175842285
+505,4.500767707824707
+506,4.5006937980651855
+507,4.500619411468506
+508,4.500545978546143
+509,4.500472068786621
+510,4.5003981590271
+511,4.500324249267578
+512,4.500250816345215
+513,4.500176906585693
+514,4.50010347366333
+515,4.500030517578125
+516,4.4999566078186035
+517,4.499882698059082
+518,4.4998087882995605
+519,4.499735355377197
+520,4.499661445617676
+521,4.499587535858154
+522,4.499513149261475
+523,4.499438762664795
+524,4.499364852905273
+525,4.499290466308594
+526,4.499216079711914
+527,4.499140739440918
+528,4.499065399169922
+529,4.498990058898926
+530,4.498913764953613
+531,4.498837947845459
+532,4.498762607574463
+533,4.49868631362915
+534,4.4986114501953125
+535,4.498536109924316
+536,4.498461723327637
+537,4.498386859893799
+538,4.498311996459961
+539,4.4982380867004395
+540,4.498164653778076
+541,4.498090744018555
+542,4.498019218444824
+543,4.497945308685303
+544,4.497872829437256
+545,4.497801303863525
+546,4.497729301452637
+547,4.49765682220459
+548,4.497586250305176
+549,4.49751615524292
+550,4.497445106506348
+551,4.49737548828125
+552,4.497305870056152
+553,4.49723482131958
+554,4.497164726257324
+555,4.497094631195068
+556,4.497025012969971
+557,4.496954441070557
+558,4.496883392333984
+559,4.496815204620361
+560,4.4967451095581055
+561,4.496673107147217
+562,4.496603488922119
+563,4.4965338706970215
+564,4.496462821960449
+565,4.496394157409668
+566,4.496323585510254
+567,4.496251583099365
+568,4.496183395385742
+569,4.49611234664917
+570,4.496041774749756
+571,4.4959716796875
+572,4.495903491973877
+573,4.4958343505859375
+574,4.49576473236084
+575,4.495695114135742
+576,4.4956278800964355
+577,4.4955573081970215
+578,4.495490074157715
+579,4.495420932769775
+580,4.495352745056152
+581,4.495282173156738
+582,4.495216369628906
+583,4.495147228240967
+584,4.495075225830078
+585,4.4950079917907715
+586,4.494940757751465
+587,4.494872093200684
+588,4.494802951812744
+589,4.494732856750488
+590,4.494668483734131
+591,4.494600296020508
+592,4.494530200958252
+593,4.494461536407471
+594,4.494396209716797
+595,4.49432897567749
+596,4.494260787963867
+597,4.494192123413086
+598,4.4941229820251465
+599,4.494059085845947
+600,4.493993759155273
+601,4.493923664093018
+602,4.493853569030762
+603,4.493789196014404
+604,4.493722915649414
+605,4.493655681610107
+606,4.493587970733643

training/normalized/average_normalized/linear/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/average_normalized/quadratic/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/average_normalized_trainer.py

@@ -0,0 +1,155 @@
+import torch
+import torch.optim as optim
+from torchdiffeq import odeint
+import numpy as np
+import os
+import shutil
+import csv
+import inspect
+
+from PendulumController import PendulumController
+from PendulumDynamics import PendulumDynamics
+
+# Device setup
+device = torch.device("cpu")
+
+# Initial conditions (theta0, omega0, alpha0, desired_theta)
+from initial_conditions import initial_conditions
+state_0 = torch.tensor(initial_conditions, dtype=torch.float32, device=device)
+
+# Device setup
+device = torch.device("cpu")
+
+# Constants
+m = 10.0
+g = 9.81
+R = 1.0
+
+# Time grid
+t_start, t_end, t_points = 0, 10, 1000
+t_span = torch.linspace(t_start, t_end, t_points, device=device)
+
+# Specify directory for storing results
+output_dir = "average_normalized"
+os.makedirs(output_dir, exist_ok=True)
+
+# Use a previously generated random seed
+random_seed = 4529
+
+# Set the seeds for reproducibility
+torch.manual_seed(random_seed)
+np.random.seed(random_seed)
+
+# Print the chosen random seed
+print(f"Random seed for torch and numpy: {random_seed}")
+
+# Initialize controller and dynamics
+controller = PendulumController().to(device)
+pendulum_dynamics = PendulumDynamics(controller, m, R, g).to(device)
+
+# Optimizer setup
+learning_rate = 1e-1
+weight_decay = 1e-4
+optimizer = optim.Adam(controller.parameters(), lr=learning_rate, weight_decay=weight_decay)
+
+# Training parameters
+num_epochs = 1001
+
+# Define loss functions
+def make_loss_fn(weight_fn):
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+    return loss_fn
+
+# Define and store weight functions with descriptions, normalized by average weight
+weight_functions = {
+    'constant': {
+        'function': lambda t: torch.ones_like(t) / torch.ones_like(t).mean(),
+        'description': 'Constant weight: All weights are 1, normalized by the average (remains 1)'
+    },
+    'linear': {
+        'function': lambda t: (t / t.max()) / (t / t.max()).mean(),
+        'description': 'Linear weight: Weights increase linearly from 0 to 1, normalized by the average weight'
+    },
+    'quadratic': {
+        'function': lambda t: ((t / t.max()) ** 2) / ((t / t.max()) ** 2).mean(),
+        'description': 'Quadratic weight: Weights increase quadratically from 0 to 1, normalized by the average weight'
+    },
+    'exponential': {
+        'function': lambda t: (torch.exp(t / t.max() * 2)) / (torch.exp(t / t.max() * 2)).mean(),
+        'description': 'Exponential weight: Weights increase exponentially, normalized by the average weight'
+    },
+    'inverse': {
+        'function': lambda t: (1 / (t / t.max() + 1)) / (1 / (t / t.max() + 1)).mean(),
+        'description': 'Inverse weight: Weights decrease inversely, normalized by the average weight'
+    },
+    'inverse_squared': {
+        'function': lambda t: (1 / ((t / t.max() + 1) ** 2)) / (1 / ((t / t.max() + 1) ** 2)).mean(),
+        'description': 'Inverse squared weight: Weights decrease inversely squared, normalized by the average weight'
+    }
+}
+
+# Training loop for each weight function
+for name, weight_info in weight_functions.items():
+    controller = PendulumController().to(device)
+    pendulum_dynamics = PendulumDynamics(controller, m, R, g).to(device)
+    optimizer = optim.Adam(controller.parameters(), lr=learning_rate, weight_decay=weight_decay)
+    loss_fn = make_loss_fn(weight_info['function'])
+
+    # File paths
+    function_output_dir = os.path.join(output_dir, name)
+    controllers_dir = os.path.join(function_output_dir, "controllers")
+
+    # Check if controllers directory exists and remove it
+    if os.path.exists(controllers_dir):
+        shutil.rmtree(controllers_dir)
+    os.makedirs(controllers_dir, exist_ok=True)
+
+    config_file = os.path.join(function_output_dir, "training_config.txt")
+    log_file = os.path.join(function_output_dir, "training_log.csv")
+
+    # Overwrite configuration and log files
+    with open(config_file, "w") as f:
+        f.write(f"Random Seed: {random_seed}\n")
+        f.write(f"Time Span: {t_start} to {t_end}, Points: {t_points}\n")
+        f.write(f"Learning Rate: {learning_rate}\n")
+        f.write(f"Weight Decay: {weight_decay}\n")
+        f.write("\nLoss Function:\n")
+        f.write(inspect.getsource(loss_fn))
+        f.write("\nTraining Cases:\n")
+        f.write("[theta0, omega0, alpha0, desired_theta]\n")
+        for case in state_0.cpu().numpy():
+            f.write(f"{case.tolist()}\n")
+
+    with open(log_file, "w", newline="") as csvfile:
+        csv_writer = csv.writer(csvfile)
+        csv_writer.writerow(["Epoch", "Loss"])
+
+    # Training loop
+    for epoch in range(num_epochs):
+        optimizer.zero_grad()
+        state_traj = odeint(pendulum_dynamics, state_0, t_span, method='rk4')
+        loss = loss_fn(state_traj, t_span)
+        loss.backward()
+        optimizer.step()
+
+        # Logging
+        with open(log_file, "a", newline="") as csvfile:
+            csv_writer = csv.writer(csvfile)
+            csv_writer.writerow([epoch, loss.item()])
+
+        # Save the model
+        model_file = os.path.join(controllers_dir, f"controller_{epoch}.pth")
+        torch.save(controller.state_dict(), model_file)
+        print(f"{model_file} saved with loss: {loss}")
+
+print("Training complete. Models and logs are saved under respective directories for each loss function.")

training/normalized/initial_conditions.py

@@ -0,0 +1,27 @@
+import torch
+from torch import pi
+
+initial_conditions = [
+    [1/6 * pi, 0.0, 0.0, 0.0],
+    [-1/6 * pi, 0.0, 0.0, 0.0],
+    [2/3 * pi, 0.0, 0.0, 0.0],
+    [-2/3 * pi, 0.0, 0.0, 0.0],
+    [0.0, 1/3 * pi, 0.0, 0.0],
+    [0.0, -1/3 * pi, 0.0, 0.0],
+    [0.0, 2 * pi, 0.0, 0.0],
+    [0.0, -2 * pi, 0.0, 0.0],
+    [0.0, 0.0, 0.0, 2 * pi],
+    [0.0, 0.0, 0.0, -2 * pi],
+    [0.0, 0.0, 0.0, 1/2 * pi],
+    [0.0, 0.0, 0.0, -1/2 * pi],
+    [0.0, 0.0, 0.0, 1/3 * pi],
+    [0.0, 0.0, 0.0, -1/3 * pi],
+    [1/4 * pi, 1 * pi, 0.0, 0.0],
+    [-1/4 * pi, -1 * pi, 0.0, 0.0],
+    [1/2 * pi, -1 * pi, 0.0, 1/3 * pi],
+    [-1/2 * pi, 1 * pi, 0.0, -1/3 * pi],
+    [1/4 * pi, 1 * pi, 0.0, 2 * pi],
+    [-1/4 * pi, -1 * pi, 0.0, 2 * pi],
+    [1/2 * pi, -1 * pi, 0.0, 4 * pi],
+    [-1/2 * pi, 1 * pi, 0.0, -4 * pi],
+]

training/normalized/max_normalized/constant/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/max_normalized/exponential/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/max_normalized/inverse/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/max_normalized/inverse_squared/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/max_normalized/linear/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/max_normalized/quadratic/training_config.txt

@@ -0,0 +1,41 @@
+Random Seed: 4529
+Time Span: 0 to 10, Points: 1000
+Learning Rate: 0.1
+Weight Decay: 0.0001
+
+Loss Function:
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+Training Cases:
+[theta0, omega0, alpha0, desired_theta]
+[0.5235987901687622, 0.0, 0.0, 0.0]
+[-0.5235987901687622, 0.0, 0.0, 0.0]
+[2.094395160675049, 0.0, 0.0, 0.0]
+[-2.094395160675049, 0.0, 0.0, 0.0]
+[0.0, 1.0471975803375244, 0.0, 0.0]
+[0.0, -1.0471975803375244, 0.0, 0.0]
+[0.0, 6.2831854820251465, 0.0, 0.0]
+[0.0, -6.2831854820251465, 0.0, 0.0]
+[0.0, 0.0, 0.0, 6.2831854820251465]
+[0.0, 0.0, 0.0, -6.2831854820251465]
+[0.0, 0.0, 0.0, 1.5707963705062866]
+[0.0, 0.0, 0.0, -1.5707963705062866]
+[0.0, 0.0, 0.0, 1.0471975803375244]
+[0.0, 0.0, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 0.0]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 0.0]
+[1.5707963705062866, -3.1415927410125732, 0.0, 1.0471975803375244]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -1.0471975803375244]
+[0.7853981852531433, 3.1415927410125732, 0.0, 6.2831854820251465]
+[-0.7853981852531433, -3.1415927410125732, 0.0, 6.2831854820251465]
+[1.5707963705062866, -3.1415927410125732, 0.0, 12.566370964050293]
+[-1.5707963705062866, 3.1415927410125732, 0.0, -12.566370964050293]

training/normalized/max_normalized_trainer.py

@@ -0,0 +1,155 @@
+import torch
+import torch.optim as optim
+from torchdiffeq import odeint
+import numpy as np
+import os
+import shutil
+import csv
+import inspect
+
+from PendulumController import PendulumController
+from PendulumDynamics import PendulumDynamics
+
+# Device setup
+device = torch.device("cpu")
+
+# Initial conditions (theta0, omega0, alpha0, desired_theta)
+from initial_conditions import initial_conditions
+state_0 = torch.tensor(initial_conditions, dtype=torch.float32, device=device)
+
+# Device setup
+device = torch.device("cpu")
+
+# Constants
+m = 10.0
+g = 9.81
+R = 1.0
+
+# Time grid
+t_start, t_end, t_points = 0, 10, 1000
+t_span = torch.linspace(t_start, t_end, t_points, device=device)
+
+# Specify directory for storing results
+output_dir = "training"
+os.makedirs(output_dir, exist_ok=True)
+
+# Use a previously generated random seed
+random_seed = 4529
+
+# Set the seeds for reproducibility
+torch.manual_seed(random_seed)
+np.random.seed(random_seed)
+
+# Print the chosen random seed
+print(f"Random seed for torch and numpy: {random_seed}")
+
+# Initialize controller and dynamics
+controller = PendulumController().to(device)
+pendulum_dynamics = PendulumDynamics(controller, m, R, g).to(device)
+
+# Optimizer setup
+learning_rate = 1e-1
+weight_decay = 1e-4
+optimizer = optim.Adam(controller.parameters(), lr=learning_rate, weight_decay=weight_decay)
+
+# Training parameters
+num_epochs = 1000
+
+# Define loss functions
+def make_loss_fn(weight_fn):
+    def loss_fn(state_traj, t_span):
+        theta = state_traj[:, :, 0]            # Size: [batch_size, t_points]
+        desired_theta = state_traj[:, :, 3]    # Size: [batch_size, t_points]
+        weights = weight_fn(t_span)            # Initially Size: [t_points]
+
+        # Reshape or expand weights to match theta dimensions
+        weights = weights.view(-1, 1)  # Now Size: [batch_size, t_points]
+
+        # Calculate the weighted loss
+        return torch.mean(weights * (theta - desired_theta) ** 2)
+
+    return loss_fn
+
+# Define and store weight functions with descriptions
+weight_functions = {
+    'constant': {
+        'function': lambda t: torch.ones_like(t),
+        'description': 'Constant weight: All weights are 1'
+    },
+    'linear': {
+        'function': lambda t: (t / t.max()) / (t / t.max()).max(),
+        'description': 'Linear weight: Weights increase linearly from 0 to 1, normalized by max'
+    },
+    'quadratic': {
+        'function': lambda t: ((t / t.max()) ** 2) / ((t / t.max()) ** 2).max(),
+        'description': 'Quadratic weight: Weights increase quadratically from 0 to 1, normalized by max'
+    },
+    'exponential': {
+        'function': lambda t: (torch.exp(t / t.max() * 2)) / (torch.exp(t / t.max() * 2)).max(),
+        'description': 'Exponential weight: Weights increase exponentially, normalized by max'
+    },
+    'inverse': {
+        'function': lambda t: (1 / (t / t.max() + 1)) / (1 / (t / t.max() + 1)).max(),
+        'description': 'Inverse weight: Weights decrease inversely, normalized by max'
+    },
+    'inverse_squared': {
+        'function': lambda t: (1 / ((t / t.max() + 1) ** 2)) / (1 / ((t / t.max() + 1) ** 2)).max(),
+        'description': 'Inverse squared weight: Weights decrease inversely squared, normalized by max'
+    }
+}
+
+# Training loop for each weight function
+for name, weight_info in weight_functions.items():
+    controller = PendulumController().to(device)
+    pendulum_dynamics = PendulumDynamics(controller, m, R, g).to(device)
+    optimizer = optim.Adam(controller.parameters(), lr=learning_rate, weight_decay=weight_decay)
+    loss_fn = make_loss_fn(weight_info['function'])
+
+    # File paths
+    function_output_dir = os.path.join(output_dir, name)
+    controllers_dir = os.path.join(function_output_dir, "controllers")
+
+    # Check if controllers directory exists and remove it
+    if os.path.exists(controllers_dir):
+        shutil.rmtree(controllers_dir)
+    os.makedirs(controllers_dir, exist_ok=True)
+
+    config_file = os.path.join(function_output_dir, "training_config.txt")
+    log_file = os.path.join(function_output_dir, "training_log.csv")
+
+    # Overwrite configuration and log files
+    with open(config_file, "w") as f:
+        f.write(f"Random Seed: {random_seed}\n")
+        f.write(f"Time Span: {t_start} to {t_end}, Points: {t_points}\n")
+        f.write(f"Learning Rate: {learning_rate}\n")
+        f.write(f"Weight Decay: {weight_decay}\n")
+        f.write("\nLoss Function:\n")
+        f.write(inspect.getsource(loss_fn))
+        f.write("\nTraining Cases:\n")
+        f.write("[theta0, omega0, alpha0, desired_theta]\n")
+        for case in state_0.cpu().numpy():
+            f.write(f"{case.tolist()}\n")
+
+    with open(log_file, "w", newline="") as csvfile:
+        csv_writer = csv.writer(csvfile)
+        csv_writer.writerow(["Epoch", "Loss"])
+
+    # Training loop
+    for epoch in range(num_epochs):
+        optimizer.zero_grad()
+        state_traj = odeint(pendulum_dynamics, state_0, t_span, method='rk4')
+        loss = loss_fn(state_traj, t_span)
+        loss.backward()
+        optimizer.step()
+
+        # Logging
+        with open(log_file, "a", newline="") as csvfile:
+            csv_writer = csv.writer(csvfile)
+            csv_writer.writerow([epoch, loss.item()])
+
+        # Save the model
+        model_file = os.path.join(controllers_dir, f"controller_{epoch}.pth")
+        torch.save(controller.state_dict(), model_file)
+        print(f"{model_file} saved with loss: {loss}")
+
+print("Training complete. Models and logs are saved under respective directories for each loss function.")