diff --git a/8382/Ershov/lb/3/main.py b/8382/Ershov/lb/3/main.py
new file mode 100644
index 00000000..e414f172
--- /dev/null
+++ b/8382/Ershov/lb/3/main.py
@@ -0,0 +1,83 @@
+import numpy as np
+from tensorflow.keras.layers import Dense
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.datasets import boston_housing
+import matplotlib.pyplot as plt
+
+
+def build_model():
+    model = Sequential()
+    model.add(Dense(64, activation='relu', input_shape=(train_data.shape[1],)))
+    model.add(Dense(64, activation='relu'))
+    model.add(Dense(1))
+    model.compile(optimizer='rmsprop', loss='mse', metrics=['mae'])
+    return model
+
+
+(train_data, train_targets), (test_data, test_targets) = boston_housing.load_data()
+
+mean = train_data.mean(axis=0)
+train_data -= mean
+std = train_data.std(axis=0)
+train_data /= std
+
+test_data -= mean
+test_data /= std
+
+k = 6
+num_val_samples = len(train_data) // k
+num_epochs = 18
+all_scores = []
+
+mean_loss = []
+mean_mae = []
+mean_val_loss = []
+mean_val_mae = []
+
+for i in range(k):
+    print('processing fold #', i)
+    val_data = train_data[i * num_val_samples: (i + 1) * num_val_samples]
+    val_targets = train_targets[i * num_val_samples: (i + 1) * num_val_samples]
+    partial_train_data = np.concatenate([train_data[:i * num_val_samples], train_data[(i + 1) * num_val_samples:]],
+                                        axis=0)
+    partial_train_targets = np.concatenate(
+        [train_targets[:i * num_val_samples], train_targets[(i + 1) * num_val_samples:]], axis=0)
+    model = build_model()
+    H = model.fit(partial_train_data, partial_train_targets, epochs=num_epochs, batch_size=1,
+                        validation_data=(val_data, val_targets), verbose=0)
+    mean_val_mae.append(H.history['val_mae'])
+    mean_mae.append(H.history['mae'])
+    plt.plot(H.history['mae'], 'g')
+    plt.plot(H.history['val_mae'], 'b')
+    plt.title('Mean absolute error' + ', i = ' + str(i + 1))
+    plt.ylabel('mae')
+    plt.xlabel('Epochs')
+    plt.legend(['Training', 'Validation'], loc='upper left')
+    plt.show()
+
+    mean_val_loss.append(H.history['val_loss'])
+    mean_loss.append(H.history['loss'])
+
+    plt.plot(H.history['loss'], 'g')
+    plt.plot(H.history['val_loss'], 'b')
+    plt.title('Model loss' + ', i = ' + str(i + 1))
+    plt.ylabel('loss')
+    plt.xlabel('Epochs')
+    plt.legend(['Training', 'Validation'], loc='upper left')
+    plt.show()
+
+plt.plot(np.mean(mean_mae, axis=0), 'g')
+plt.plot(np.mean(mean_val_mae, axis=0), 'b')
+plt.title('Mean model mae')
+plt.ylabel('mae')
+plt.xlabel('Epochs')
+plt.legend(['Training', 'Validation'], loc='upper left')
+plt.show()
+
+plt.plot(np.mean(mean_loss, axis=0), 'g')
+plt.plot(np.mean(mean_val_loss, axis=0), 'b')
+plt.title('Mean model loss')
+plt.ylabel('loss')
+plt.xlabel('Epochs')
+plt.legend(['Training', 'Validation'], loc='upper left')
+plt.show()
\ No newline at end of file
diff --git a/8382/Ershov/lb/3/report.pdf b/8382/Ershov/lb/3/report.pdf
new file mode 100644
index 00000000..7d9964e5
Binary files /dev/null and b/8382/Ershov/lb/3/report.pdf differ