Update app.py

app.py

@@ -1,30 +1,71 @@
-import gradio as gr
-from fastai.vision.all import *
+# Import necessary libraries
+import gradio as gr  # Gradio for creating web interfaces
+from fastai.vision.all import *  # FastAI library for deep learning tasks
 
-# Define the GrayscaleTransform class
+# Define a custom transformation class for converting images to grayscale
 class GrayscaleTransform(Transform):
+    """
+    Custom transformation class to convert images to grayscale.
+    This is used to ensure that the input images match the format
+    used during model training.
+    """
     def encodes(self, img: PILImage):
-        return img.convert("L")
+        """
+        Convert the input image to grayscale.
+        
+        Args:
+            img (PILImage): The input image in PIL format.
+        
+        Returns:
+            PIL.Image: The grayscale version of the input image.
+        """
+        return img.convert("L")  # 'L' mode represents grayscale images
 
-# Load the exported model
+# Load the pre-trained model
 learn = load_learner('clocker.pkl')
+"""
+load_learner function loads a saved FastAI learner object.
+The 'clocker.pkl' file contains the trained model, including
+its architecture, weights, and any necessary preprocessing steps.
+"""
 
 def classify_image(img):
+    """
+    Classify the input image using the loaded model.
+    
+    Args:
+        img: The input image to be classified.
+    
+    Returns:
+        dict: A dictionary containing the prediction probabilities for each class.
+    """
+    # Make a prediction using the loaded model
     pred, _, probs = learn.predict(img)
+    # Return a dictionary with class probabilities
     return {
-        "average woman": float(probs[0]),
-        "transgender woman": float(probs[1])
+        "average woman": float(probs[0]),  # Probability for "average woman" class
+        "transgender woman": float(probs[1])  # Probability for "transgender woman" class
     }
 
 # Create the Gradio interface
 iface = gr.Interface(
-    fn=classify_image,
-    inputs=gr.Image(),
-    outputs=gr.Label(num_top_classes=2),
-    title="Transfem Clocker AI",
-    description="Upload an image of a woman and this will guess if she is trans.",
-   
+    fn=classify_image,  # The function to be called when the interface is used
+    inputs=gr.Image(),  # Input component: an image upload widget
+    outputs=gr.Label(num_top_classes=2),  # Output component: label with top 2 classes
+    title="Transfem Clocker AI",  # Title of the web interface
+    description="Upload an image of a woman and this will guess if she is trans.",  # Description of the interface
 )
+"""
+gr.Interface creates a web interface for the model:
+- fn: The function to be called when an image is uploaded
+- inputs: Specifies that the input should be an image
+- outputs: Displays the top 2 class probabilities as labels
+- title and description: Provides context for users
+"""
 
-# Launch the interface
-iface.launch()
+# Launches the interface
+iface.launch()
+"""
+This starts the Gradio interface, making it accessible via a web browser.
+it is my first ever AI web app!
+"""