from tensorflow.keras import layers, Model, Input
from tensorflow.keras.layers import Rescaling

input_layer = Input(shape=(256, 256, 3))
x = Rescaling(1.0 / 255)(input_layer)
x = layers.Conv2D(32, (3, 3), activation='relu')(input_layer)
x = layers.MaxPooling2D((2, 2))(x)
x = layers.Conv2D(64, (3, 3), activation='relu')(x)
x = layers.MaxPooling2D((2, 2))(x)
x = layers.Conv2D(128, (3, 3), activation='relu')(x)
x = layers.MaxPooling2D((2, 2))(x)
x = layers.Conv2D(256, (3, 3), activation='relu')(x)
x = layers.MaxPooling2D((2, 2))(x)
x = layers.GlobalAveragePooling2D(name='global_avg_pool')(x)
x = layers.Dense(128, activation='relu')(x)
x = layers.Dropout(0.2)(x)
output_layer = layers.Dense(39, activation='softmax')(x)

model = Model(inputs=input_layer, outputs=output_layer)
model.compile(
    optimizer='adam',
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy']
)

callbacks = [
    tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5),
    tf.keras.callbacks.ModelCheckpoint('best_model.h5', save_best_only=True)
]

history = model.fit(
    train_dataset,
    epochs=30,
    validation_data=val_dataset,
    callbacks=callbacks,
    verbose=1
)

from tensorflow.keras.applications import InceptionV3
from tensorflow.keras.layers import Dense, GlobalAveragePooling2D, Dropout
from tensorflow.keras.models import Model
import tensorflow as tf

# Load model InceptionV3 tanpa output layer
base_model = InceptionV3(weights='imagenet', include_top=False, input_shape=(224, 224, 3))
base_model.trainable = False  # Freeze base model

# Dapatkan lapisan hingga mixed7
x = base_model.get_layer('mixed7').output

# Tambahkan custom classifier
x = GlobalAveragePooling2D()(x)
x = Dense(128, activation='relu')(x)
x = Dropout(0.2)(x)
predictions = Dense(num_classes, activation='softmax')(x)

# Buat model akhir
model = Model(inputs=base_model.input, outputs=predictions)
model.summary()