Update app.py

app.py

@@ -1,27 +1,26 @@
 ### app.py
-import gradio as gr
 import torch
-from loaded_model import CNNModel
-import loaded_model
-
-model = CNNModel()
-model.load_state_dict(torch.load(f="https://huggingface.co/spaces/gpbhupinder/test/blob/main/model_-%2023%20june%202024%2019_22.pt"))
+import requests
+from PIL import Image
+from torchvision import transforms
+import gradio as gr
+model = torch.hub.load('https://huggingface.co/spaces/gpbhupinder/test/blob/main/model_-%2023%20june%202024%2019_22.pt', pretrained=True).eval()
 
 
-# Define a function to make predictions with your model
-def classify_image(image):
-    # Preprocess the image
-    preprocess = loaded_model.create_transformer()
-    
-    image_tensor = preprocess(image)
-    image_tensor = image_tensor.unsqueeze(0)
+# Download human-readable labels for ImageNet.
+response = requests.get("https://git.io/JJkYN")
+labels = response.text.split("\n")
 
-    # Make prediction
-    with torch.no_grad():
-        output = model(image_tensor)
-    _, predicted_class = torch.max(output, 1)
+def predict(inp):
+  inp = transforms.ToTensor()(inp).unsqueeze(0)
+  with torch.no_grad():
+    prediction = torch.nn.functional.softmax(model(inp)[0], dim=0)
+    confidences = {labels[i]: float(prediction[i]) for i in range(1000)}
+  return confidences
 
-    return f"Predicted class: {predicted_class.item()}"
 
 
-gr.Interface(fn=classify_image, inputs="sketchpad", outputs="label").launch()
+gr.Interface(fn=predict,
+             inputs=gr.Image(type="pil"),
+             outputs=gr.Label(num_top_classes=3),
+             examples=["lion.jpg", "cheetah.jpg"]).launch()