Upload folder using huggingface_hub

best_model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1758608733d41d3eb3874611d7638d2cdaca9068380b851b4b984e56787bb072
+size 17620960

class_names.txt

@@ -0,0 +1,45 @@
+airplane
+airport
+baseball_diamond
+basketball_court
+beach
+bridge
+chaparral
+church
+circular_farmland
+cloud
+commercial_area
+dense_residential
+desert
+forest
+freeway
+golf_course
+ground_track_field
+harbor
+industrial_area
+intersection
+island
+lake
+meadow
+medium_residential
+mobile_home_park
+mountain
+overpass
+palace
+parking_lot
+railway
+railway_station
+rectangular_farmland
+river
+roundabout
+runway
+sea_ice
+ship
+snowberg
+sparse_residential
+stadium
+storage_tank
+tennis_court
+terrace
+thermal_power_station
+wetland

prediction.py

@@ -0,0 +1,26 @@
+#Code for checking 1 sample image 
+import os
+import numpy as np
+from tensorflow.keras.preprocessing import image
+from tensorflow.keras.models import load_model
+
+# Set correct image size based on training
+IMG_SIZE = (128, 128)
+
+# Load your model
+model = load_model("best_model.h5")
+
+# Get class names from folders
+DATA_DIR = r"C:\Users\arsul\Desktop\RESISC45\NWPU-RESISC45"
+CLASS_NAMES = sorted(os.listdir(DATA_DIR))
+
+# Load and preprocess the image
+img_path = r"C:\Users\arsul\Desktop\RESISC45\ball-courts.jpg"
+img = image.load_img(img_path, target_size=IMG_SIZE)
+img_array = image.img_to_array(img)
+img_array = np.expand_dims(img_array, axis=0)  # Add batch dimension
+
+# Make prediction
+preds = model.predict(img_array)
+predicted_class = CLASS_NAMES[np.argmax(preds)]
+print(f"Predicted class: {predicted_class}")

train_resisc45.py

@@ -0,0 +1,117 @@
+import tensorflow as tf
+from tensorflow.keras.applications import EfficientNetB0
+from tensorflow.keras.layers import Dense, GlobalAveragePooling2D, Dropout
+from tensorflow.keras.models import Model
+from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint, ReduceLROnPlateau
+import matplotlib.pyplot as plt
+import os
+
+# Paths
+DATASET_DIR = r"C:\Path\To\NWPU-RESISC45"  # Path to your Dataset
+IMG_SIZE = (128, 128)
+BATCH_SIZE = 16
+NUM_CLASSES = 45
+EPOCHS = 50
+SEED = 42
+
+# Sanity check: verify class folders
+class_names = sorted(os.listdir(DATASET_DIR))
+print(f"Detected {len(class_names)} classes:", class_names)
+
+# Augmentation for training set
+data_augmentation = tf.keras.Sequential([
+    tf.keras.layers.RandomFlip("horizontal_and_vertical"),
+    tf.keras.layers.RandomRotation(0.2),
+    tf.keras.layers.RandomZoom(0.1),
+])
+
+# Load Datasets
+train_ds = tf.keras.preprocessing.image_dataset_from_directory(
+    DATASET_DIR,
+    validation_split=0.2,
+    subset="training",
+    seed=SEED,
+    image_size=IMG_SIZE,
+    batch_size=BATCH_SIZE,
+    label_mode='categorical'
+)
+
+val_ds = tf.keras.preprocessing.image_dataset_from_directory(
+    DATASET_DIR,
+    validation_split=0.2,
+    subset="validation",
+    seed=SEED,
+    image_size=IMG_SIZE,
+    batch_size=BATCH_SIZE,
+    label_mode='categorical'
+)
+
+# Apply augmentation only to training
+train_ds = train_ds.map(lambda x, y: (data_augmentation(x, training=True), y))
+
+# Prefetch
+AUTOTUNE = tf.data.AUTOTUNE
+train_ds = train_ds.prefetch(buffer_size=AUTOTUNE)
+val_ds = val_ds.prefetch(buffer_size=AUTOTUNE)
+
+# Build Model
+base_model = EfficientNetB0(weights='imagenet', include_top=False, input_shape=(*IMG_SIZE, 3))
+base_model.trainable = False  # Freeze base model initially
+
+x = base_model.output
+x = GlobalAveragePooling2D()(x)
+x = Dropout(0.3)(x)
+output = Dense(NUM_CLASSES, activation='softmax')(x)
+
+model = Model(inputs=base_model.input, outputs=output)
+
+model.compile(optimizer='adam',
+              loss='categorical_crossentropy',
+              metrics=['accuracy'])
+
+model.summary()
+
+# Callbacks
+callbacks = [
+    ModelCheckpoint("best_model.h5", save_best_only=True, monitor="val_accuracy", mode="max"),
+    EarlyStopping(patience=10, restore_best_weights=True),
+    ReduceLROnPlateau(factor=0.2, patience=5, min_lr=1e-6)
+]
+
+# Train
+history = model.fit(
+    train_ds,
+    validation_data=val_ds,
+    epochs=EPOCHS,
+    callbacks=callbacks
+)
+
+# Optional: Fine-tune after warmup (unfreeze base model)
+base_model.trainable = True
+model.compile(optimizer=tf.keras.optimizers.Adam(1e-5),  # Lower LR for fine-tuning
+              loss='categorical_crossentropy',
+              metrics=['accuracy'])
+
+history_finetune = model.fit(
+    train_ds,
+    validation_data=val_ds,
+    epochs=10,
+    callbacks=callbacks
+)
+
+# Plotting
+plt.plot(history.history["accuracy"] + history_finetune.history["accuracy"], label="train acc")
+plt.plot(history.history["val_accuracy"] + history_finetune.history["val_accuracy"], label="val acc")
+plt.xlabel("Epochs")
+plt.ylabel("Accuracy")
+plt.legend()
+plt.title("Training vs Validation Accuracy")
+plt.grid()
+plt.savefig("training_accuracy.png")
+plt.show()
+
+# Save class names
+with open("class_names.txt", "w") as f:
+    for name in class_names:
+        f.write(name + "\n")
+