Update app.py

app.py

@@ -1,17 +1,17 @@
-import gradio as gr
+]import gradio as gr
 from PIL import Image
 import torch
 from torchvision import models, transforms
 import PyPDF2  # For reading patient reports (PDFs)
 
-# Load the pre-trained model (ResNet18 with new weight parameter)
-model = models.resnet18(weights="IMAGENET1K_V1")  # Updated to the new weight parameter in torchvision 0.13+
-model.eval()
+# Load a more specialized pre-trained model (DenseNet, similar to CheXNet)
+model = models.densenet121(pretrained=True)  # DenseNet is good for image classification tasks like chest X-rays
+model.eval()  # Set the model to evaluation mode
 
 # Define image preprocessing function
 def preprocess_image(image):
     transform = transforms.Compose([
-        transforms.Resize((224, 224)),
+        transforms.Resize((224, 224)),  # Resize to fit the model input size
         transforms.ToTensor(),
     ])
     return transform(image).unsqueeze(0)
@@ -27,7 +27,7 @@ def predict_xray(image):
     conditions = ["Normal", "Pneumonia", "Cancer", "TB", "Other"]
     results = {conditions[i]: float(probs[i]) for i in range(len(conditions))}
     
-    # Determine the most likely condition and provide detailed feedback
+    # Identify the most likely condition and calculate the confidence
     most_likely_condition = max(results, key=results.get)
     confidence = results[most_likely_condition] * 100  # Convert to percentage
     
@@ -64,7 +64,7 @@ def predict_xray(image):
         detailed_results += f"<li><b>{condition}:</b> {prob*100:.2f}%</li>"
     detailed_results += "</ul>"
     
-    # Additional advice based on the most likely condition
+    # Additional feedback for the most likely condition
     additional_feedback = condition_details.get(most_likely_condition, "Please consult with a doctor for further details.")
     
     return summary, detailed_results, additional_feedback