Rename app to app.py

app

@@ -1,40 +0,0 @@
-import gradio as gr
-from PIL import Image
-import torch
-import pickle
-import torchvision.transforms as transforms
-
-# Load the pickled model
-model_path = "birds_classifier.pkl"
-try:
-    with open(model_path, "rb") as f:
-        model = pickle.load(f)
-    model.eval()  # Set model to evaluation mode
-except Exception as e:
-    raise Exception(f"Failed to load model: {str(e)}")
-
-# Define image preprocessing (adjust these transforms based on your model's training)
-preprocess = transforms.Compose([
-    transforms.Resize((224, 224)),  # Match your model's expected input size
-    transforms.ToTensor(),
-    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),  # ImageNet defaults
-])
-
-# Replace with your actual list of bird species (in the order the model was trained)
-class_labels = ["Sparrow", "Eagle", "Blue Jay", "Cardinal"]  # Update this!
-
-# Prediction function
-def classify_bird(image):
-    try:
-        if image is None:
-            return "Please upload an image of a bird."
-        
-        # Preprocess the uploaded image
-        img = preprocess(image).unsqueeze(0)  # Add batch dimension
-        
-        # Make prediction automatically
-        with torch.no_grad():
-            outputs = model(img)  # Model outputs logits or probabilities
-        
-        # Get the predicted species
-        predicted_idx = outputs.argmax(-1).item()

app.py

@@ -0,0 +1,46 @@
+import gradio as gr
+import pickle
+import numpy as np
+from PIL import Image
+
+# Load the trained model
+with open("bird_classifier.pkl", "rb") as f:
+    model = pickle.load(f)
+
+# Get class names automatically from the model
+try:
+    class_names = model.classes_  # Works for scikit-learn models
+except AttributeError:
+    # If the model doesn't have classes_, you'd need a fallback or custom logic
+    raise ValueError("Model does not have 'classes_' attribute. Please provide class names manually or adjust the code.")
+
+# Define the prediction function
+def classify_bird(image):
+    # Preprocess the image (adjust this based on how your model was trained)
+    img = Image.fromarray(image.astype("uint8"), "RGB")  # Convert to PIL Image
+    img = img.resize((224, 224))  # Example resize, adjust to your model's input size
+    img_array = np.array(img) / 255.0  # Normalize if your model expects this
+    img_array = np.expand_dims(img_array, axis=0)  # Add batch dimension
+
+    # Make prediction
+    prediction = model.predict(img_array)  # Adjust based on your model's method
+    
+    # Handle prediction output
+    if len(prediction.shape) > 1:  # If prediction is a probability array (e.g., softmax output)
+        predicted_class = class_names[np.argmax(prediction)]
+    else:  # If prediction is a single class index (e.g., scikit-learn's default)
+        predicted_class = class_names[prediction[0]]
+    
+    return predicted_class
+
+# Create the Gradio interface
+interface = gr.Interface(
+    fn=classify_bird,                # Prediction function
+    inputs=gr.Image(type="numpy"),   # Input is an image, returned as NumPy array
+    outputs=gr.Textbox(),            # Output is text (bird species)
+    title="Bird Classifier",
+    description="Upload an image of a bird and get its species predicted!"
+)
+
+# Launch the app
+interface.launch()