src/train/trainer.py

"""train_model.py

This module trains a simple image classification model using TensorFlow/Keras.

"""

import os
import tensorflow as tf
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.preprocessing import image
import numpy as np
from scipy import misc

# Constants
IMG_SIZE = (224, 224)
BATCH_SIZE = 32
EPOCHS = 10
NUM_CLASSES = 5

def preprocess_image(img):
    """
    Preprocess the input image for model prediction.

    Parameters:
    - img: Input image.

    Returns:
    - img: Preprocessed image.
    """
    img = image.img_to_array(img)
    img = np.expand_dims(img, axis=0)
    img /= 255.0
    return img

def train_model(dataset_path):
    """
    Train the image classification model.

    Parameters:
    - dataset_path (str): Path to the dataset.

    Returns:
    - model: Trained Keras model.
    """

    # Ensure the dataset path is correct
    dataset_path = os.path.abspath(dataset_path)

    # Data preprocessing
    datagen = ImageDataGenerator(
        rescale=1./255,
        validation_split=0.2
    )

    train_generator = datagen.flow_from_directory(
        dataset_path,
        target_size=IMG_SIZE,
        batch_size=BATCH_SIZE,
        class_mode='categorical',
        subset='training'
    )

    validation_generator = datagen.flow_from_directory(
        dataset_path,
        target_size=IMG_SIZE,
        batch_size=BATCH_SIZE,
        class_mode='categorical',
        subset='validation'
    )

    # Model definition
    model = Sequential()
    model.add(Conv2D(32, (3, 3), activation='relu', input_shape=(224, 224, 3)))
    model.add(MaxPooling2D((2, 2)))
    model.add(Flatten())
    model.add(Dense(64, activation='relu'))
    model.add(Dense(NUM_CLASSES, activation='softmax'))

    # Model compilation
    model.compile(optimizer=Adam(learning_rate=0.001), loss='categorical_crossentropy', metrics=['accuracy'])

    # Model training
    history = model.fit(
        train_generator,
        epochs=EPOCHS,
        validation_data=validation_generator
    )

    return model

if __name__ == "__main__":
    script_directory = os.path.dirname(os.path.abspath(__file__))
    dataset_path = os.path.join(script_directory, '../../data')
    trained_model = train_model(dataset_path)
    trained_model.save(os.path.join(script_directory, '../../models/football_logo_model.h5'))