Update app.py

app.py

@@ -3,15 +3,14 @@ import numpy as np
 import matplotlib.pyplot as plt
 
 from tensorflow.keras.models import Sequential
-from tensorflow.keras.layers import Dense, Input
+from tensorflow.keras.layers import Dense, Input, Dropout
 from tensorflow.keras.optimizers import Adam
+from tensorflow.keras.callbacks import EarlyStopping
 
 from sklearn.datasets import make_moons, make_circles, make_blobs
 from sklearn.model_selection import train_test_split
 from sklearn.preprocessing import StandardScaler
 
-
-# Sidebar - UI Controls
 st.sidebar.title("🔧 Model Settings")
 
 with st.sidebar.expander("🧠 Dataset Settings"):
@@ -28,7 +27,11 @@ with st.sidebar.expander("⚙️ Model Hyperparameters"):
     num_neurons = st.slider("Neurons per Hidden Layer", 1, 100, 16)
     hidden_layers = st.slider("Number of Hidden Layers", 1, 5, 2)
 
-# Dataset Generator
+with st.sidebar.expander("🛠 Regularization"):
+    use_earlystop = st.checkbox("Use EarlyStopping", value=True)
+    use_dropout = st.checkbox("Use Dropout", value=True)
+    dropout_rate = st.slider("Dropout Rate", 0.0, 0.5, 0.3) if use_dropout else 0.0
+
 if dataset == "Moons":
     x, y = make_moons(n_samples=n_samples, noise=noise, random_state=42)
 elif dataset == "Circles":
@@ -41,60 +44,87 @@ x = scaler.fit_transform(x)
 
 x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=split, random_state=27)
 
-# Model Definition
-model = Sequential()
-model.add(Input(shape=(2,)))
-for _ in range(hidden_layers):
-    model.add(Dense(units=num_neurons, activation=activation))
-model.add(Dense(1, activation="sigmoid"))
-
-model.compile(optimizer=Adam(learning_rate=lr), loss='binary_crossentropy', metrics=['accuracy'])
-
-# Training
-history = model.fit(x_train, y_train,
-                    validation_data=(x_test, y_test),
-                    batch_size=batch,
-                    epochs=epochs,
-                    verbose=0)
-
-test_loss, test_acc = model.evaluate(x_test, y_test, verbose=0)
-
-# UI Display
-st.title("🧪 Neural Network Playground")
-st.markdown("Visualize how a simple feedforward neural network learns to classify synthetic datasets.")
-
-st.subheader("📈 Model Performance")
-st.success(f"**Test Accuracy:** {test_acc:.4f}")
-st.info(f"**Test Loss:** {test_loss:.4f}")
-
-# Decision Boundary Plot
-st.subheader("🔍 Decision Boundary")
-
-x_min, x_max = x[:, 0].min() - 1, x[:, 0].max() + 1
-y_min, y_max = x[:, 1].min() - 1, x[:, 1].max() + 1
-xx, yy = np.meshgrid(np.linspace(x_min, x_max, 300),
-                     np.linspace(y_min, y_max, 300))
-grid = np.c_[xx.ravel(), yy.ravel()]
-preds = model.predict(grid, verbose=0).reshape(xx.shape)
-
-fig, ax = plt.subplots(figsize=(7, 6))
-ax.contourf(xx, yy, preds, cmap='RdBu', alpha=0.6)
-ax.scatter(x[:, 0], x[:, 1], c=y, cmap='RdBu', edgecolors='k', s=30)
-ax.set_title("Decision Boundary")
-ax.set_xlabel("Feature 1")
-ax.set_ylabel("Feature 2")
-st.pyplot(fig)
-
-# Training History Plot
-def plot_loss(history):
+callbacks = [EarlyStopping(patience=10, restore_best_weights=True)] if use_earlystop else []
+
+def build_model(with_dropout=False):
+    model = Sequential()
+    model.add(Input(shape=(2,)))
+    for _ in range(hidden_layers):
+        model.add(Dense(units=num_neurons, activation=activation))
+        if with_dropout:
+            model.add(Dropout(dropout_rate))
+    model.add(Dense(1, activation="sigmoid"))
+    model.compile(optimizer=Adam(learning_rate=lr), loss='binary_crossentropy', metrics=['accuracy'])
+    return model
+
+base_model = build_model(False)
+dropout_model = build_model(True)
+
+base_hist = base_model.fit(x_train, y_train,
+                           validation_data=(x_test, y_test),
+                           batch_size=batch,
+                           epochs=epochs,
+                           callbacks=[],
+                           verbose=0)
+
+early_model = build_model(False)
+early_hist = early_model.fit(x_train, y_train,
+                             validation_data=(x_test, y_test),
+                             batch_size=batch,
+                             epochs=epochs,
+                             callbacks=callbacks,
+                             verbose=0)
+
+dropout_hist = dropout_model.fit(x_train, y_train,
+                                 validation_data=(x_test, y_test),
+                                 batch_size=batch,
+                                 epochs=epochs,
+                                 callbacks=callbacks,
+                                 verbose=0)
+
+def plot_decision_boundary(model, title):
+    x_min, x_max = x[:, 0].min() - 1, x[:, 0].max() + 1
+    y_min, y_max = x[:, 1].min() - 1, x[:, 1].max() + 1
+    xx, yy = np.meshgrid(np.linspace(x_min, x_max, 300),
+                         np.linspace(y_min, y_max, 300))
+    grid = np.c_[xx.ravel(), yy.ravel()]
+    preds = model.predict(grid, verbose=0).reshape(xx.shape)
+
+    fig, ax = plt.subplots()
+    ax.contourf(xx, yy, preds, cmap='RdBu', alpha=0.6)
+    ax.scatter(x[:, 0], x[:, 1], c=y, cmap='RdBu', edgecolors='k', s=25)
+    ax.set_title(title)
+    ax.set_xlabel("Feature 1")
+    ax.set_ylabel("Feature 2")
+    return fig
+
+def plot_loss(history, title):
     fig, ax = plt.subplots()
     ax.plot(history.history['loss'], label='Train Loss')
-    ax.plot(history.history['val_loss'], label='Validation Loss')
-    ax.set_title("Training vs Validation Loss")
-    ax.set_xlabel("Epoch")
+    ax.plot(history.history['val_loss'], label='Val Loss')
+    ax.set_title(title)
+    ax.set_xlabel("Epochs")
     ax.set_ylabel("Loss")
     ax.legend()
     return fig
 
-st.subheader("📉 Training Loss")
-st.pyplot(plot_loss(history))
+st.title("🧪 Regularization Comparison")
+
+st.markdown("Below are the decision boundaries and loss curves for three configurations:")
+
+cols = st.columns(3)
+
+with cols[0]:
+    st.markdown("### 🔹 Base Model")
+    st.pyplot(plot_decision_boundary(base_model, "Base Model"))
+    st.pyplot(plot_loss(base_hist, "Base Model Loss"))
+
+with cols[1]:
+    st.markdown("### 🟢 With EarlyStopping")
+    st.pyplot(plot_decision_boundary(early_model, "EarlyStopping"))
+    st.pyplot(plot_loss(early_hist, "EarlyStopping Loss"))
+
+with cols[2]:
+    st.markdown("### 🔸 With Dropout")
+    st.pyplot(plot_decision_boundary(dropout_model, "Dropout Model"))
+    st.pyplot(plot_loss(dropout_hist, "Dropout Loss"))