Update

augment.py

@@ -12,22 +12,22 @@ for task in ["1"]:
         if class_label != ".DS_Store":
             print("Augmenting images in class: ", class_label, " in Task ", task)
             # Create a temp folder to combine the raw data and the external data
-            if not os.path.exists(f"{TEMP_DATA_DIR}Task {task}/{class_label}/"):
-                os.makedirs(f"{TEMP_DATA_DIR}Task {task}/{class_label}/")
-            if os.path.exists(f"{RAW_DATA_DIR}Task {task}/{class_label}"):
-                for file in os.listdir(f"{RAW_DATA_DIR}Task {task}/{class_label}"):
+            if not os.path.exists(f"{TEMP_DATA_DIR}{task}/{class_label}/"):
+                os.makedirs(f"{TEMP_DATA_DIR}{task}/{class_label}/")
+            if os.path.exists(f"{RAW_DATA_DIR}{task}/{class_label}"):
+                for file in os.listdir(f"{RAW_DATA_DIR}{task}/{class_label}"):
                     shutil.copy(
-                        f"{RAW_DATA_DIR}Task {task}/{class_label}/{file}",
-                        f"{TEMP_DATA_DIR}Task {task}/{class_label}/{str(uuid.uuid4())}.png",
+                        f"{RAW_DATA_DIR}{task}/{class_label}/{file}",
+                        f"{TEMP_DATA_DIR}{task}/{class_label}/{str(uuid.uuid4())}.png",
                     )
-            if os.path.exists(f"{EXTERNAL_DATA_DIR}Task {task}/{class_label}"):
-                for file in os.listdir(f"{EXTERNAL_DATA_DIR}Task {task}/{class_label}"):
+            if os.path.exists(f"{EXTERNAL_DATA_DIR}{task}/{class_label}"):
+                for file in os.listdir(f"{EXTERNAL_DATA_DIR}{task}/{class_label}"):
                     shutil.copy(
-                        f"{EXTERNAL_DATA_DIR}Task {task}/{class_label}/{file}",
-                        f"{TEMP_DATA_DIR}Task {task}/{class_label}/{str(uuid.uuid4())}.png",
+                        f"{EXTERNAL_DATA_DIR}{task}/{class_label}/{file}",
+                        f"{TEMP_DATA_DIR}{task}/{class_label}/{str(uuid.uuid4())}.png",
                     )
             p = Augmentor.Pipeline(
-                f"{TEMP_DATA_DIR}Task {task}/{class_label}",
+                f"{TEMP_DATA_DIR}{task}/{class_label}",
                 output_directory=f"{class_label}/",
                 save_format="png",
             )
@@ -42,20 +42,20 @@ for task in ["1"]:
             p.sample(100 - len(p.augmentor_images))
             # Move the folder to data/train/Task 1/augmented
             # Create the folder if it does not exist
-            if not os.path.exists(f"{AUG_DATA_DIR}Task {task}/"):
-                os.makedirs(f"{AUG_DATA_DIR}Task {task}/")
+            if not os.path.exists(f"{AUG_DATA_DIR}{task}/"):
+                os.makedirs(f"{AUG_DATA_DIR}{task}/")
             # Move all images in the data/train/Task 1/i folder to data/train/Task 1/augmented/i
             os.rename(
-                f"{TEMP_DATA_DIR}Task {task}/{class_label}/{class_label}",
-                f"{AUG_DATA_DIR}Task {task}/{class_label}",
+                f"{TEMP_DATA_DIR}{task}/{class_label}/{class_label}",
+                f"{AUG_DATA_DIR}{task}/{class_label}",
             )
             # Rename all the augmented images to [01, 02, 03]
             number = 0
-            for file in os.listdir(f"{AUG_DATA_DIR}Task {task}/{class_label}"):
+            for file in os.listdir(f"{AUG_DATA_DIR}{task}/{class_label}"):
                 number = int(number) + 1
                 if len(str(number)) == 1:
                     number = "0" + str(number)
                 os.rename(
-                    f"{AUG_DATA_DIR}Task {task}/{class_label}/{file}",
-                    f"{AUG_DATA_DIR}Task {task}/{class_label}/{number}.png",
+                    f"{AUG_DATA_DIR}{task}/{class_label}/{file}",
+                    f"{AUG_DATA_DIR}{task}/{class_label}/{number}.png",
                 )

configs.py

@@ -6,11 +6,11 @@ from models import *
 
 # Constants
 RANDOM_SEED = 123
-BATCH_SIZE = 16
+BATCH_SIZE = 32
 NUM_EPOCHS = 100
-LEARNING_RATE = 5.847227637580824e-05
+LEARNING_RATE = 0.00017588413773574044
 STEP_SIZE = 10
-GAMMA = 1.0
+GAMMA = 0.3
 DEVICE = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 NUM_PRINT = 100
 TASK = 1
@@ -22,7 +22,7 @@ NUM_CLASSES = 7
 EARLY_STOPPING_PATIENCE = 20
 CLASSES = ['Alzheimer Disease', 'Cerebral Palsy', 'Dystonia', 'Essential Tremor', 'Healthy', 'Huntington Disease', 'Parkinson Disease']
 MODEL_SAVE_PATH = "output/checkpoints/model.pth"
-MODEL = efficientnet_b1(num_classes=NUM_CLASSES)
+MODEL = squeezenet1_0(num_classes=NUM_CLASSES)
 
 print(CLASSES)
 

tuning.py

@@ -12,7 +12,7 @@ from torch.utils.tensorboard import SummaryWriter
 DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 EPOCHS = 10
 N_TRIALS = 50
-TIMEOUT = 3600  # 1 hour
+TIMEOUT = None
 
 # Create a TensorBoard writer
 writer = SummaryWriter(log_dir="output/tensorboard/tuning")
@@ -116,8 +116,8 @@ if __name__ == "__main__":
 
     # Optimize the hyperparameters
     study.optimize(
-        objective, n_trials=100, timeout=3600
-    )  # Adjust the number of trials and timeout as needed
+        objective, n_trials=N_TRIALS, timeout=TIMEOUT
+    )
 
     # Print the best trial
     best_trial = study.best_trial