Update pages/1_Random_Data.py

pages/1_Random_Data.py

@@ -9,6 +9,13 @@ from sklearn.preprocessing import StandardScaler
 st.title("🔢 Random Data Playground")
 st.markdown("Define hyperparameters, generate synthetic data, and train your neural network interactively.")
 
+# Initialize session state variables
+if 'train_requested' not in st.session_state:
+    st.session_state.train_requested = False
+if 'data_ready' not in st.session_state:
+    st.session_state.data_ready = False
+
+# Form for user input
 with st.form("data_form"):
     st.subheader("🔧 Model Configuration")
     col1, col2 = st.columns(2)
@@ -39,7 +46,27 @@ if submitted:
     X_scaled = scaler.fit_transform(X)
     X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.3, random_state=42)
 
-    # Decision Region Before Training
+    # Save to session state
+    st.session_state.X_scaled = X_scaled
+    st.session_state.y = y
+    st.session_state.X_train = X_train
+    st.session_state.X_test = X_test
+    st.session_state.y_train = y_train
+    st.session_state.y_test = y_test
+    st.session_state.data_ready = True
+    st.session_state.hyperparams = {
+        'batch_size': batch_size,
+        'epochs': epochs,
+        'learning_rate': learning_rate,
+        'n_layers': n_layers,
+        'n_neurons': n_neurons
+    }
+
+# Visualize data
+if st.session_state.data_ready:
+    X_scaled = st.session_state.X_scaled
+    y = st.session_state.y
+
     st.subheader("🌀 Generated Data & Decision Region (Before Training)")
     h = 0.02
     x_min, x_max = X_scaled[:, 0].min() - 1, X_scaled[:, 0].max() + 1
@@ -51,57 +78,60 @@ if submitted:
     st.pyplot(fig)
 
     if st.button("🚀 Train Model"):
-        progress_bar = st.progress(0)
-        status_text = st.empty()
-
-        # Build model
-        model = tf.keras.Sequential()
-        model.add(tf.keras.layers.Input(shape=(2,)))
-        for _ in range(n_layers):
-            model.add(tf.keras.layers.Dense(n_neurons, activation='relu'))
-        model.add(tf.keras.layers.Dense(1, activation='sigmoid'))
-
-        optimizer = tf.keras.optimizers.Adam(learning_rate=learning_rate)
-        model.compile(optimizer=optimizer, loss='binary_crossentropy', metrics=['accuracy'])
-
-        # Custom training loop to show progress
-        class StreamlitCallback(tf.keras.callbacks.Callback):
-            def on_epoch_end(self, epoch, logs=None):
-                percent = (epoch + 1) / epochs
-                progress_bar.progress(min(percent, 1.0))
-                status_text.text(f"Epoch {epoch+1}/{epochs} - Loss: {logs['loss']:.4f}, Val Loss: {logs['val_loss']:.4f}")
-
-        history = model.fit(
-            X_train, y_train,
-            batch_size=batch_size,
-            epochs=epochs,
-            validation_split=0.2,
-            verbose=0,
-            callbacks=[StreamlitCallback()]
-        )
-
-        st.success("✅ Model Trained!")
-
-        # Decision Region After Training
-        st.subheader("🌈 Decision Region After Training")
-        preds = model.predict(np.c_[xx.ravel(), yy.ravel()]).reshape(xx.shape)
-        fig, ax = plt.subplots()
-        ax.contourf(xx, yy, preds, cmap=plt.cm.RdBu, alpha=0.6)
-        ax.scatter(X_scaled[:, 0], X_scaled[:, 1], c=y, cmap=plt.cm.bwr, edgecolor='k')
-        st.pyplot(fig)
-
-        # Loss Plot
-        st.subheader("📉 Training vs Validation Loss")
-        fig, ax = plt.subplots()
-        ax.plot(history.history['loss'], label='Train Loss')
-        ax.plot(history.history['val_loss'], label='Validation Loss')
-        ax.set_xlabel("Epochs")
-        ax.set_ylabel("Loss")
-        ax.legend()
-        st.pyplot(fig)
-
-        # Loss values
-        final_train_loss = history.history['loss'][-1]
-        final_val_loss = history.history['val_loss'][-1]
-        st.write(f"🧮 **Final Train Loss**: `{final_train_loss:.4f}`")
-        st.write(f"🧪 **Final Validation Loss**: `{final_val_loss:.4f}`")
+        st.session_state.train_requested = True
+
+# Train the model
+if st.session_state.get("train_requested", False):
+    st.subheader("🧠 Training Model")
+    X_train = st.session_state.X_train
+    y_train = st.session_state.y_train
+    X_test = st.session_state.X_test
+    y_test = st.session_state.y_test
+    params = st.session_state.hyperparams
+
+    progress_bar = st.progress(0)
+    status_text = st.empty()
+
+    # Callback to update progress
+    class StreamlitCallback(tf.keras.callbacks.Callback):
+        def on_epoch_end(self, epoch, logs=None):
+            percent = (epoch + 1) / params['epochs']
+            progress_bar.progress(min(percent, 1.0))
+            status_text.text(f"Epoch {epoch+1}/{params['epochs']} - Loss: {logs['loss']:.4f}, Val Loss: {logs['val_loss']:.4f}")
+
+    # Build model
+    model = tf.keras.Sequential()
+    model.add(tf.keras.layers.Input(shape=(2,)))
+    for _ in range(params['n_layers']):
+        model.add(tf.keras.layers.Dense(params['n_neurons'], activation='relu'))
+    model.add(tf.keras.layers.Dense(1, activation='sigmoid'))
+
+    optimizer = tf.keras.optimizers.Adam(learning_rate=params['learning_rate'])
+    model.compile(optimizer=optimizer, loss='binary_crossentropy', metrics=['accuracy'])
+
+    history = model.fit(X_train, y_train, batch_size=params['batch_size'], epochs=params['epochs'],
+                        validation_split=0.2, verbose=0, callbacks=[StreamlitCallback()])
+
+    st.success("✅ Model Trained!")
+
+    # Decision Region After Training
+    st.subheader("🌈 Decision Region After Training")
+    preds = model.predict(np.c_[xx.ravel(), yy.ravel()]).reshape(xx.shape)
+    fig, ax = plt.subplots()
+    ax.contourf(xx, yy, preds, cmap=plt.cm.RdBu, alpha=0.6)
+    ax.scatter(X_scaled[:, 0], X_scaled[:, 1], c=y, cmap=plt.cm.bwr, edgecolor='k')
+    st.pyplot(fig)
+
+    # Loss Plots
+    st.subheader("📈 Training vs Validation Loss")
+    fig, ax = plt.subplots()
+    ax.plot(history.history['loss'], label='Train Loss')
+    ax.plot(history.history['val_loss'], label='Validation Loss')
+    ax.set_xlabel("Epoch")
+    ax.set_ylabel("Loss")
+    ax.legend()
+    st.pyplot(fig)
+
+    # Display last values
+    st.markdown(f"**Final Train Loss:** {history.history['loss'][-1]:.4f}")
+    st.markdown(f"**Final Validation Loss:** {history.history['val_loss'][-1]:.4f}")