Update app.py

app.py

@@ -1,24 +1,14 @@
-import gradio as gr
-import matplotlib.pyplot as plt
-import numpy as np
 import os
-import PIL
+import zipfile
+import gdown
+import pathlib
 import tensorflow as tf
-
 from tensorflow import keras
 from tensorflow.keras import layers
 from tensorflow.keras.models import Sequential
-
-
-from PIL import Image
-import gdown
-import zipfile
-
-import pathlib
-
-
-
-
+import matplotlib.pyplot as plt
+import gradio as gr
+import numpy as np
 
 # Define the Google Drive shareable link
 gdrive_url = 'https://drive.google.com/file/d/1HjHYlQyRz5oWt8kehkt1TiOGRRlKFsv8/view?usp=drive_link'
@@ -48,9 +38,9 @@ except zipfile.BadZipFile:
 os.remove(local_zip_file)
 
 # Convert the extracted directory path to a pathlib.Path object
-data_dir = pathlib.Path(extracted_path)
+data_dir = pathlib.Path('extracted_files/Pest_Dataset')
 
-# Print the directory structure to debug
+# Verify the directory structure
 for root, dirs, files in os.walk(extracted_path):
     level = root.replace(extracted_path, '').count(os.sep)
     indent = ' ' * 4 * (level)
@@ -59,143 +49,143 @@ for root, dirs, files in os.walk(extracted_path):
     for f in files:
         print(f"{subindent}{f}")
 
-import pathlib
-# Path to the dataset directory
-data_dir = pathlib.Path('extracted_files/Pest_Dataset')
-data_dir = pathlib.Path(data_dir)
-
-
-bees = list(data_dir.glob('bees/*'))
-print(bees[0])
-PIL.Image.open(str(bees[0]))
-
+# Set image dimensions and batch size
+img_height, img_width = 180, 180
+batch_size = 32
 
-bees = list(data_dir.glob('bees/*'))
-print(bees[0])
-PIL.Image.open(str(bees[0]))
+# Create training and validation datasets
+train_ds = tf.keras.preprocessing.image_dataset_from_directory(
+    data_dir,
+    validation_split=0.2,
+    subset="training",
+    seed=123,
+    image_size=(img_height, img_width),
+    batch_size=batch_size
+)
 
-batch_size = 32
-img_height = 180
-img_width = 180
-
-train_ds = tf.keras.utils.image_dataset_from_directory(
-  data_dir,
-  validation_split=0.2,
-  subset="training",
-  seed=123,
-  image_size=(img_height, img_width),
-  batch_size=batch_size)
-
-val_ds = tf.keras.utils.image_dataset_from_directory(
-  data_dir,
-  validation_split=0.2,
-  subset="validation",
-  seed=123,
-  image_size=(img_height, img_width),
-  batch_size=batch_size)
+val_ds = tf.keras.preprocessing.image_dataset_from_directory(
+    data_dir,
+    validation_split=0.2,
+    subset="validation",
+    seed=123,
+    image_size=(img_height, img_width),
+    batch_size=batch_size
+)
 
 class_names = train_ds.class_names
 print(class_names)
 
-import matplotlib.pyplot as plt
-
+# Display some sample images
 plt.figure(figsize=(10, 10))
 for images, labels in train_ds.take(1):
-  for i in range(9):
-    ax = plt.subplot(3, 3, i + 1)
-    plt.imshow(images[i].numpy().astype("uint8"))
-    plt.title(class_names[labels[i]])
-    plt.axis("off")
+    for i in range(9):
+        ax = plt.subplot(3, 3, i + 1)
+        plt.imshow(images[i].numpy().astype("uint8"))
+        plt.title(class_names[labels[i]])
+        plt.axis("off")
 
+# Enhanced data augmentation
+data_augmentation = keras.Sequential(
+    [
+        layers.RandomFlip("horizontal", input_shape=(img_height, img_width, 3)),
+        layers.RandomRotation(0.2),
+        layers.RandomZoom(0.2),
+        layers.RandomContrast(0.2),
+        layers.RandomBrightness(0.2),
+    ]
+)
 
-for image_batch, labels_batch in train_ds:
-  print(image_batch.shape)
-  print(labels_batch.shape)
-  break
+# Display augmented images
+plt.figure(figsize=(10, 10))
+for images, _ in train_ds.take(1):
+    for i in range(9):
+        augmented_images = data_augmentation(images)
+        ax = plt.subplot(3, 3, i + 1)
+        plt.imshow(augmented_images[0].numpy().astype("uint8"))
+        plt.axis("off")
 
+# Define a deeper CNN model with more regularization techniques
+num_classes = len(class_names)
+model = Sequential()
 
-AUTOTUNE = tf.data.AUTOTUNE
+model.add(data_augmentation)
+model.add(layers.Rescaling(1./255))
 
-train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)
-val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)
+model.add(layers.Conv2D(32, 3, padding='same', activation='relu'))
+model.add(layers.BatchNormalization())
+model.add(layers.MaxPooling2D())
 
-normalization_layer = layers.Rescaling(1./255)
+model.add(layers.Conv2D(64, 3, padding='same', activation='relu'))
+model.add(layers.BatchNormalization())
+model.add(layers.MaxPooling2D())
 
-normalized_ds = train_ds.map(lambda x, y: (normalization_layer(x), y))
-image_batch, labels_batch = next(iter(normalized_ds))
-first_image = image_batch[0]
-# Notice the pixel values are now in `[0,1]`.
-print(np.min(first_image), np.max(first_image))
+model.add(layers.Conv2D(128, 3, padding='same', activation='relu'))
+model.add(layers.BatchNormalization())
+model.add(layers.MaxPooling2D())
 
-data_augmentation = keras.Sequential(
-  [
-    layers.RandomFlip("horizontal",
-                      input_shape=(img_height,
-                                  img_width,
-                                  3)),
-    layers.RandomRotation(0.1),
-    layers.RandomZoom(0.1),
-  ]
-)
+model.add(layers.Conv2D(256, 3, padding='same', activation='relu'))
+model.add(layers.BatchNormalization())
+model.add(layers.MaxPooling2D())
 
+model.add(layers.Conv2D(512, 3, padding='same', activation='relu'))
+model.add(layers.BatchNormalization())
+model.add(layers.MaxPooling2D())
 
-plt.figure(figsize=(10, 10))
-for images, _ in train_ds.take(1):
-  for i in range(9):
-    augmented_images = data_augmentation(images)
-    ax = plt.subplot(3, 3, i + 1)
-    plt.imshow(augmented_images[0].numpy().astype("uint8"))
-    plt.axis("off")
+model.add(layers.Dropout(0.5))
+model.add(layers.Flatten())
 
-num_classes = len(class_names)
+model.add(layers.Dense(256, activation='relu'))
+model.add(layers.Dropout(0.5))
+
+model.add(layers.Dense(num_classes, activation='softmax', name="outputs"))
 
-model = Sequential([
-  data_augmentation,
-  layers.Rescaling(1./255),
-  layers.Conv2D(16, 3, padding='same', activation='relu'),
-  layers.MaxPooling2D(),
-  layers.Conv2D(32, 3, padding='same', activation='relu'),
-  layers.MaxPooling2D(),
-  layers.Conv2D(64, 3, padding='same', activation='relu'),
-  layers.MaxPooling2D(),
-  layers.Dropout(0.2),
-  layers.Flatten(),
-  layers.Dense(128, activation='relu'),
-  layers.Dense(num_classes, name="outputs")
-])
-
-model.compile(optimizer='adam',
-              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
+model.compile(optimizer=keras.optimizers.Adam(learning_rate=1e-4),
+              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
               metrics=['accuracy'])
 
 model.summary()
 
-epochs = 15
+# Implement early stopping
+from tensorflow.keras.callbacks import EarlyStopping
+
+early_stopping = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)
+
+# Learning rate scheduler
+def scheduler(epoch, lr):
+    if epoch < 10:
+        return lr
+    else:
+        return lr * tf.math.exp(-0.1)
+
+lr_scheduler = keras.callbacks.LearningRateScheduler(scheduler)
+
+# Train the model
+epochs = 30
 history = model.fit(
-  train_ds,
-  validation_data=val_ds,
-  epochs=epochs
+    train_ds,
+    validation_data=val_ds,
+    epochs=epochs,
+    callbacks=[early_stopping, lr_scheduler]
 )
 
-import gradio as gr
-import numpy as np
-import tensorflow as tf
-
+# Define the prediction function
 def predict_image(img):
     img = np.array(img)
     img_resized = tf.image.resize(img, (180, 180))
     img_4d = tf.expand_dims(img_resized, axis=0)
     prediction = model.predict(img_4d)[0]
-    return {class_names[i]: float(prediction[i]) for i in range(len(class_names))}
-
+    predicted_class = np.argmax(prediction)
+    predicted_label = class_names[predicted_class]
+    return {predicted_label: f"{float(prediction[predicted_class]):.2f}"}
 
+# Set up Gradio interface
 image = gr.Image()
 label = gr.Label(num_top_classes=1)
 
 # Define custom CSS for background image
 custom_css = """
 body {
-    background-image: url('\extracted_files\Pest_Dataset\bees\bees (444).jpg');
+    background-image: url('extracted_files/Pest_Dataset/bees/bees (444).jpg');
     background-size: cover;
     background-repeat: no-repeat;
     background-attachment: fixed;
@@ -208,6 +198,6 @@ gr.Interface(
     inputs=image,
     outputs=label,
     title="Welcome to Agricultural Pest Image Classification",
-    description="The image data set used was obtaied from Kaggle and has a collection of 12 different types of agricultral pests: Ants, Bees, Beetles, Caterpillars, Earthworms, Earwigs, Grasshoppers, Moths, Slugs, Snails, Wasps, and Weevils",
+    description="The image data set used was obtained from Kaggle and has a collection of 12 different types of agricultural pests: Ants, Bees, Beetles, Caterpillars, Earthworms, Earwigs, Grasshoppers, Moths, Slugs, Snails, Wasps, and Weevils",
     css=custom_css
-).launch(debug=True)
+).launch(debug=True)