ahh

.gitignore

@@ -185,3 +185,6 @@ ipython_config.py
 
 
 .conda*/
+
+data/
+# for rn

load_data.py

@@ -76,6 +76,12 @@ def validate_view(data: pd.DataFrame, verbose=False):
 # if __name__ == "__main__":
 def load_this_data(filename, verbose=False):
     d = load_data(filename)
+    # if verbose:
+    #     print(d.columns)
     d = date_combine(d)
+    # if verbose:
+    #     print(d.columns)
     d = validate_view(d, verbose)
+    # if verbose:
+    #     print(d.columns)
     return d

tutorial/pt_tuto.py

@@ -16,6 +16,10 @@ from sklearn.preprocessing import MinMaxScaler
 from torch import nn, optim
 from torch.utils.data import DataLoader, TensorDataset
 
+import sys
+
+sys.path.append(r"c:\Users\Admin\Documents\lecture ass shit\ml1\codemidterm\ml_midterm")
+
 from load_data import load_this_data
 
 # Step 2: Data Preparation
@@ -66,25 +70,30 @@ def create_sequences(data, n_lookback):
     return np.array(x), np.array(y)
 
 
-# Split into training and testing sets (80% train, 20% test)
+# Split into training and testing sets
-train_data, test_data = train_test_split(data_scaled, test_size=0.2, shuffle=False)
+split_index = int(0.8 * len(data_scaled))
+train_data = data_scaled[:split_index]
+test_data = data_scaled[split_index:]
 
 # Create sequences for training and testing
 train_x, train_y = create_sequences(train_data, n_lookback)
 test_x, test_y = create_sequences(test_data, n_lookback)
 
+# train_y = scaler.transform(train_y)
+# test_y = scaler.transform(test_y)
 # Convert sequences to PyTorch tensors
 train_x = torch.tensor(train_x, dtype=torch.float32)
 train_y = torch.tensor(train_y, dtype=torch.float32)
 test_x = torch.tensor(test_x, dtype=torch.float32)
 test_y = torch.tensor(test_y, dtype=torch.float32)
 
+
 # Create datasets and dataloaders
 train_dataset = TensorDataset(train_x, train_y)
 test_dataset = TensorDataset(test_x, test_y)
 
-train_loader = DataLoader(train_dataset, batch_size=128, shuffle=True)
+train_loader = DataLoader(train_dataset, batch_size=32, shuffle=True)
-test_loader = DataLoader(test_dataset, batch_size=128, shuffle=False)
+test_loader = DataLoader(test_dataset, batch_size=32, shuffle=False)
 
 # Step 3: Define the LSTM Model
 # -----------------------------
@@ -92,43 +101,74 @@ test_loader = DataLoader(test_dataset, batch_size=128, shuffle=False)
 # The forward method specifies how data flows through the model.
 
 
+# class LSTMModel(nn.Module):
+#     def __init__(self, n_features, n_lookback, n_neurons):
+#         """
+#         Initialize the LSTM model.
+
+#         Args:
+#             n_features: Number of input features (e.g., 1 for univariate time series).
+#             n_lookback: Number of past timesteps to consider.
+#             n_neurons: Number of neurons in each LSTM layer.
+#         """
+#         super(LSTMModel, self).__init__()
+#         self.lstm1 = nn.LSTM(
+#             n_features,
+#             n_neurons,
+#             batch_first=True,
+#             dropout=0.2,
+#         )
+#         self.lstm2 = nn.LSTM(
+#             n_neurons,
+#             n_neurons,
+#             batch_first=True,
+#         )
+#         self.lstm3 = nn.LSTM(n_neurons, n_neurons, batch_first=True)
+#         self.lstm4 = nn.LSTM(n_neurons, n_neurons, batch_first=True)
+#         self.fc = nn.Linear(n_neurons, 1)
+#         self.dropout = nn.Dropout(0.2)
+
+#     def forward(self, x):
+#         """
+#         Forward pass through the model.
+
+#         Args:
+#             x: Input tensor of shape (batch, seq_len, n_features).
+
+#         Returns:
+#             Output tensor of shape (batch, 1).
+#         """
+#         x, _ = self.lstm1(x)
+#         x, _ = self.lstm2(x)
+#         x, _ = self.lstm3(x)
+#         x, _ = self.lstm4(x)
+#         x = self.dropout(x)
+#         x = self.fc(x[:, -1, :])  # Use the last timestep's output
+#         return x
+
+print("Input sequences (x):", train_x.shape)
+print("Target values (y):", train_y.shape)
+
+print("Sample input:", train_x[0])
+print("Corresponding target:", train_y[0])
+
 class LSTMModel(nn.Module):
     def __init__(self, n_features, n_lookback, n_neurons):
-        """
-        Initialize the LSTM model.
-
-        Args:
-            n_features: Number of input features (e.g., 1 for univariate time series).
-            n_lookback: Number of past timesteps to consider.
-            n_neurons: Number of neurons in each LSTM layer.
-        """
         super(LSTMModel, self).__init__()
-        self.lstm = nn.LSTM(
+        self.lstm = nn.LSTM(n_features, n_neurons, batch_first=True, dropout=0.2)
-            input_size=n_features,  # Number of input features
+        self.lstm2 = nn.LSTM(n_neurons, n_neurons, batch_first=True, dropout=0.2)
-            hidden_size=n_neurons,  # Number of neurons in the LSTM
+
-            num_layers=2,  # Number of LSTM layers
+        self.fc = nn.Linear(n_neurons, 1)
-            batch_first=True,  # Input shape is (batch, seq, feature)
-            dropout=0.2,  # Dropout for regularization
-        )
-        self.fc = nn.Linear(n_neurons, 1)  # Fully connected layer for output
 
     def forward(self, x):
-        """
+        x, _ = self.lstm(x)
-        Forward pass through the model.
+        x, _ = self.lstm2(x)
-
+        x = self.fc(x[:, -1, :])  # Use the last timestep's output
-        Args:
+        return x
-            x: Input tensor of shape (batch, seq_len, n_features).
-
-        Returns:
-            Output tensor of shape (batch, 1).
-        """
-        lstm_out, _ = self.lstm(x)  # LSTM outputs (ignore hidden states)
-        out = self.fc(lstm_out[:, -1, :])  # Use the last timestep's output
-        return out
 
 
 # Instantiate the model
-n_neurons = 64  # Number of neurons in the LSTM
+n_neurons = 128  # Number of neurons in the LSTM
 model = LSTMModel(n_features, n_lookback, n_neurons)
 
 # Step 4: Define Loss Function and Optimizer
@@ -136,14 +176,14 @@ model = LSTMModel(n_features, n_lookback, n_neurons)
 # Loss function: Mean Squared Error (MSE) for regression tasks.
 # Optimizer: Adam optimizer for efficient training.
 criterion = nn.MSELoss()
-optimizer = optim.Adam(model.parameters(), lr=0.0001)
+optimizer = optim.Adam(model.parameters(), lr=0.001)
 
 # Step 5: Train the Model
 # -----------------------
 # PyTorch requires an explicit training loop. This gives you full control
 # over the training process.
 
-epochs = 100  # Number of training epochs
+epochs = 50  # Number of training epochs
 
 # Check for CUDA and set the device
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
@@ -152,6 +192,9 @@ print(f"Using device: {device}")
 # Move the model to the device
 model = LSTMModel(n_features, n_lookback, n_neurons).to(device)
 
+# scheduler = optim.lr_scheduler.ReduceLROnPlateau(
+    # optimizer, mode="min", patience=5, factor=0.5, min_lr=1e-6
+# )
 # Move data to the device during training and evaluation
 for epoch in range(epochs):
     model.train()  # Set the model to training mode
@@ -163,23 +206,39 @@ for epoch in range(epochs):
         y_batch = y_batch.to(device)
         # Reshape input to (batch, seq_len, n_features)
         x_batch = x_batch.view(-1, n_lookback, n_features)
-        y_batch = y_batch.view(-1, 1)
+        # y_batch = y_batch.view(-1, 1)
-
+        # print("x_batch shape:", x_batch.shape)
+        # print("y_batch shape:", y_batch.shape)
+        # print("x_batch dtype:", x_batch.dtype)
+        # print("y_batch dtype:", y_batch.dtype)  
+        # break
         # Zero the gradients from the previous step
         optimizer.zero_grad()
 
         # Forward pass: Compute predictions
         predictions = model(x_batch)
-
+        # if epoch == 32:
+            # for name, param in model.named_parameters():
+            #     if param.grad is not None:
+            #         print(f"{name} gradient mean: {param.grad.mean().item()}")
+            #     else:
+            #         print(f"{name} has no gradient")
+            # print("Predictions device:", predictions.device)
+            # print("y_batch device:", y_batch.device)
+            # print("Predictions shape:", predictions.shape)
+            # print("y_batch shape:", y_batch.shape)
         # Compute the loss
         loss = criterion(predictions, y_batch)
 
         # Backward pass: Compute gradients
         loss.backward()
 
+        # Clip gradients to prevent exploding gradients
+        torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=1.0)
+
         # Update model parameters
         optimizer.step()
-
+        # scheduler.step(train_loss / len(train_loader))
         # Accumulate the loss for reporting
         train_loss += loss.item()
 
@@ -187,8 +246,9 @@ for epoch in range(epochs):
     print(f"Epoch {epoch+1}/{epochs}, Loss: {train_loss/len(train_loader):.4f}")
 
     if epoch % 10 == 0:  # Print every 10 epochs
-        print(f"Sample prediction: {predictions[0].item()}, Actual: {y_batch[0].item()}")
+        print(
-
+            f"Sample prediction: {predictions[0].item()}, Actual: {y_batch[0].item()}"
+        )
 # Step 6: Evaluate the Model
 # --------------------------
 # Switch to evaluation mode and disable gradient computation.
@@ -203,7 +263,7 @@ with torch.no_grad():
         y_batch = y_batch.to(device)
 
         # Reshape input to (batch, seq_len, n_features)
-        x_batch = x_batch.view(-1, n_lookback, n_features)
+        # x_batch = x_batch.view(-1, n_lookback, n_features)
 
         # Make predictions
         preds = model(x_batch)
@@ -223,9 +283,22 @@ predictions = scaler.inverse_transform(predictions)
 actuals = scaler.inverse_transform(actuals)
 
 # Plot the results
+print("Sample scaled train_y:", train_y[:5])
+print("Sample scaled test_y:", test_y[:5])
+
+print("Predictions shape:", predictions.shape)
+print("Actuals shape:", actuals.shape)
+
+print("Sample inverse-transformed predictions:", predictions[:5])
+print("Sample inverse-transformed actuals:", actuals[:5])
 
 plt.figure(figsize=(20, 5))
 plt.plot(actuals, label="Actual")
 plt.plot(predictions, label="Predicted")
 plt.legend()
 plt.show()
+
+from sklearn.metrics import r2_score, mean_squared_error
+
+print("R² Score:", r2_score(actuals, predictions))
+print("MSE:", mean_squared_error(actuals, predictions))

tutorial/pytorch_test.py

@@ -1,7 +1,124 @@
-""" this uses new tech
+"""this uses new tech
 make by conda new -p path python=3.12; conda activate -p path; conda install all those shi
 
 conda install pytorch torchvision torchaudio pytorch-cuda -c pytorch -c nvidia
 """
 
+import sys
 
+sys.path.append(".")
+from load_data import load_this_data
+import torch
+
+# a = load_this_data("dataset.csv")
+
+x = torch.tensor([[1, 2], [3, 4]])
+print(x)
+
+
+print(x.shape)       # Output: torch.Size([2, 2])
+print(x.dtype)       # Output: torch.int64
+print(x.device)      # Output: cpu
+
+y = torch.ones_like(x)
+print(x + y)
+print(x * y)
+
+
+x = torch.randn(3, requires_grad=True)
+y = x * 2
+z = y.mean()
+z.backward()
+print(x.grad)
+
+import torch.nn as nn
+
+class SimpleNet(nn.Module):
+    def __init__(self):
+        super(SimpleNet, self).__init__()
+        self.fc = nn.Linear(2, 1)
+
+    def forward(self, x):
+        return self.fc(x)
+
+
+model = SimpleNet()
+sample_input = torch.tensor([[1.0, 2.0]])
+output = model(sample_input)
+print(output)
+
+import torch.optim as optim
+
+criterion = nn.MSELoss()
+optimizer = optim.SGD(model.parameters(), lr=0.01)
+
+for epoch in range(100):
+    optimizer.zero_grad()
+    outputs = model(sample_input)
+    loss = criterion(outputs, torch.tensor([[1.0]]))
+    loss.backward()
+    optimizer.step()
+    if (epoch+1) % 10 == 0:
+        print(f'Epoch [{epoch+1}/100], Loss: {loss.item():.4f}')
+
+model.eval()
+
+sample_input_2 = torch.tensor([[1.0, 2.0], [3.0, 4.0]])
+output_2 = model(sample_input_2)
+print(output_2)
+
+
+from torchvision import datasets, transforms
+from torch.utils.data import DataLoader
+
+# Define a transform to normalize the data
+transform = transforms.Compose([transforms.ToTensor()])
+
+# Download and load the training data
+train_dataset = datasets.MNIST(root='data', train=True, download=True, transform=transform)
+train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)
+
+
+# import torch.nn as nn
+import torch.nn.functional as F
+
+class Net(nn.Module):
+    def __init__(self):
+        super(Net, self).__init__()
+        self.fc1 = nn.Linear(28 * 28, 128)
+        self.fc2 = nn.Linear(128, 10)
+
+    def forward(self, x):
+        x = x.view(-1, 28 * 28)  # Flatten the input
+        x = F.relu(self.fc1(x))
+        x = self.fc2(x)
+        return x
+
+model_2 = Net()
+criterion_2 = nn.CrossEntropyLoss()
+optimizer_2 = torch.optim.SGD(model.parameters(), lr=0.01)
+
+
+for epoch in range(35):  # loop over the dataset multiple times
+    for inputs, labels in train_loader:
+        optimizer_2.zero_grad()           # zero the parameter gradients
+        outputs = model_2(inputs)         # forward
+        loss = criterion_2(outputs, labels)  # compute loss
+        loss.backward()                 # backward
+        optimizer_2.step()                # optimize
+    print(f'Epoch {epoch+1}, Loss: {loss.item()}')
+
+
+test_dataset = datasets.MNIST(root='data', train=False, download=True, transform=transform)
+test_loader = DataLoader(test_dataset, batch_size=64, shuffle=False)
+
+
+correct = 0
+total = 0
+with torch.no_grad():
+    for inputs, labels in test_loader:
+        outputs = model_2(inputs)
+        _, predicted = torch.max(outputs.data, 1)
+        total += labels.size(0)
+        correct += (predicted == labels).sum().item()
+print(f'Accuracy: {100 * correct / total}%')