I am having issues with transfer learning the Mobilenetv2 for image classification of food items. I had initially used resnet but was facing the same overfitting issue but mobilenet runs faster so i decided to stick with that. I’ve posted code snippits below:
RELEVENT CODE:
# Create an ImageDataGenerator with data augmentation for training
data_generator = keras.preprocessing.image.ImageDataGenerator(
preprocessing_function=keras.applications.mobilenet.preprocess_input,
# rescale=1./255,
validation_split=0.3,
rotation_range=30, # Random rotations
width_shift_range=0.2, # Horizontal shifts
height_shift_range=0.2, # Vertical shifts
shear_range=0.2, # Shear transformations
zoom_range=0.2, # Zoom
horizontal_flip=True, # Horizontal flips
)
# Load and preprocess training data
train_data_flow = data_generator.flow_from_directory(
dataset_path,
target_size=(128, 128), # Resize images to 224x224
batch_size=32,
class_mode='categorical',
subset='training' # Use training subset
)
# Load and preprocess validation data
val_data_flow = data_generator.flow_from_directory(
dataset_path,
target_size=(128, 128), # Resize images to 224x224
batch_size=32,
class_mode='categorical',
subset='validation' # Use validation subset
)
# Load the model from TensorFlow Hub
model_url = "https://www.kaggle.com/models/google/mobilenet-v2/TensorFlow2/035-128-classification/2"
hub_layer = hub.KerasLayer(model_url, input_shape=(128, 128, 3) , trainable=True, arguments=dict(batch_norm_momentum=0.997))
Creating the model:
# Create a Sequential model with dropout and batch normalization
model = keras.Sequential([
hub_layer,
layers.Dropout(0.2), # Lower dropout rate
layers.Dense(256, activation='relu'),
layers.BatchNormalization(), # Batch normalization
layers.Dropout(0.2), # Dropout
layers.Dense(9, activation='softmax')
])
# Build the Sequential model
model.build((None, 128, 128, 3))
# Summary of the model
model.summary()
Fitting the model:
# Define the EarlyStopping callback
early_stopping_callback = keras.callbacks.EarlyStopping(
monitor='val_loss', # Monitor validation loss
patience=20, # Number of epochs with no improvement after which training will be stopped
restore_best_weights=True # Restore model weights from the epoch with the best value of the monitored quantity
)
# Fit the model with early stopping
history = model.fit(
train_data_flow,
validation_data=val_data_flow,
epochs=100,
callbacks=[early_stopping_callback], # Pass the EarlyStopping callback as a list
shuffle= True
)
Typical case of overfitting. I’ve tried data augmentation, different splits for training and validation , shuffling etc. Any Ideas?
Here are some of the epochs:
Epoch 1/100
113/113 [==============================] - 73s 641ms/step - loss: 0.5302 - accuracy: 0.8623 - val_loss: 3.2597 - val_accuracy: 0.4723
Epoch 2/100
113/113 [==============================] - 70s 618ms/step - loss: 0.4798 - accuracy: 0.8751 - val_loss: 2.4527 - val_accuracy: 0.6070
Epoch 3/100
113/113 [==============================] - 70s 621ms/step - loss: 0.4891 - accuracy: 0.8695 - val_loss: 2.0448 - val_accuracy: 0.6018
Epoch 4/100
113/113 [==============================] - 70s 624ms/step - loss: 0.4618 - accuracy: 0.8851 - val_loss: 2.0285 - val_accuracy: 0.5764
Epoch 5/100
113/113 [==============================] - 70s 618ms/step - loss: 0.4320 - accuracy: 0.8948 - val_loss: 1.7363 - val_accuracy: 0.6344
Epoch 6/100
113/113 [==============================] - 70s 619ms/step - loss: 0.3853 - accuracy: 0.9051 - val_loss: 1.9907 - val_accuracy: 0.5901
Epoch 7/100
113/113 [==============================] - 70s 618ms/step - loss: 0.4188 - accuracy: 0.8909 - val_loss: 2.0851 - val_accuracy: 0.5777
Epoch 8/100
113/113 [==============================] - 71s 625ms/step - loss: 0.3877 - accuracy: 0.9062 - val_loss: 1.4624 - val_accuracy: 0.6630
Epoch 9/100
113/113 [==============================] - 70s 621ms/step - loss: 0.3589 - accuracy: 0.9115 - val_loss: 1.6237 - val_accuracy: 0.6376
Epoch 10/100
113/113 [==============================] - 70s 618ms/step - loss: 0.3782 - accuracy: 0.8998 - val_loss: 1.6615 - val_accuracy: 0.6207
Epoch 11/100
113/113 [==============================] - 70s 621ms/step - loss: 0.3714 - accuracy: 0.9040 - val_loss: 2.5165 - val_accuracy: 0.4782
Epoch 12/100
113/113 [==============================] - 71s 629ms/step - loss: 0.3514 - accuracy: 0.9149 - val_loss: 1.9595 - val_accuracy: 0.5901
Epoch 13/100
...
113/113 [==============================] - 76s 669ms/step - loss: 0.2445 - accuracy: 0.9466 - val_loss: 1.3978 - val_accuracy: 0.6630
Epoch 44/100
113/113 [==============================] - 71s 632ms/step - loss: 0.2272 - accuracy: 0.9527 - val_loss: 1.3801 - val_accuracy: 0.6760
FOR PREDICITONS:
from sklearn.metrics import classification_report
predictions = model.predict(val_data_flow)
# Convert predictions to class labels
predicted_classes = np.argmax(predictions, axis=1)
# Step 4: Evaluate the Model
# Get the true labels from the test data
true_labels = val_data_flow.classes
# Calculate accuracy
accuracy = np.mean(predicted_classes == true_labels)
# Accuracy as the score
print("Model Accuracy on Test Set:", accuracy)
# sklearn's classification_report for more detailed metrics
print("Classification Report:")
print(classification_report(true_labels, predicted_classes, target_names=list(val_data_flow.class_indices.keys())))
Results:
Model Accuracy on Test Set: 0.19388418998048146
Classification Report:
precision recall f1-score support
1 0.32 0.30 0.31 495
2 0.00 0.00 0.00 21
3 0.17 0.21 0.19 199
4 0.13 0.09 0.11 154
5 0.00 0.00 0.00 8
6 0.11 0.06 0.08 32
7 0.11 0.07 0.08 199
8 0.16 0.19 0.17 231
9 0.13 0.18 0.15 198
accuracy 0.19 1537
macro avg 0.12 0.12 0.12 1537
weighted avg 0.19 0.19 0.19 1537