Upload 3 files

app.py

@@ -0,0 +1,76 @@
+import gradio as gr
+import torch
+import torch.nn as nn
+from torchvision import transforms, models
+from PIL import Image
+
+# ===============================
+# 📌 Classes (same order as ImageFolder during training)
+# Replace with trainers.classes if available
+# ===============================
+class_names = [
+    "Acura", "Alfa Romeo", "Aston Martin", "Audi", "Bentley",
+    "BMW", "Bugatti", "Buick", "Cadillac", "Chevrolet",
+    "Chrysler", "Citroen", "Dodge", "Ferrari", "Fiat",
+    "Ford", "Genesis", "GMC", "Honda", "Hyundai",
+    "Infiniti", "Jaguar", "Jeep", "Kia", "Lamborghini",
+    "Land Rover", "Lexus", "Lincoln", "Maserati", "Mazda",
+    "McLaren", "Mercedes", "Mini", "Mitsubishi", "Nissan",
+    "Pagani", "Peugeot", "Porsche", "Ram", "Renault",
+    "Rolls Royce", "Saab", "Subaru", "Suzuki", "Tesla",
+    "Toyota", "Volkswagen", "Volvo", "Others1", "Others2"
+]
+
+# ===============================
+# 📌 Inference Transform (NO augmentation)
+# ===============================
+transform = transforms.Compose([
+    transforms.Lambda(lambda x: x.convert("RGB")),
+    transforms.Resize((224,224)),
+    transforms.ToTensor(),
+    transforms.Normalize([0.5]*3, [0.5]*3)
+])
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+
+# ===============================
+# 📌 Load Model
+# ===============================
+def load_model(weights_path="best_model.pth"):
+    base_model = models.resnet50(weights=models.ResNet50_Weights.DEFAULT)
+    in_features = base_model.fc.in_features
+    base_model.fc = nn.Sequential(
+        nn.Linear(in_features, 512),
+        nn.ReLU(),
+        nn.Dropout(0.5),
+        nn.Linear(512, len(class_names))
+    )
+    base_model.load_state_dict(torch.load(weights_path, map_location=device))
+    return base_model.to(device).eval()
+
+model = load_model()
+
+# ===============================
+# 📌 Prediction Function
+# ===============================
+def predict(image):
+    img = transform(image).unsqueeze(0).to(device)
+    with torch.no_grad():
+        outputs = model(img)
+        probs = torch.softmax(outputs, dim=1)[0]
+        top5_probs, top5_idx = torch.topk(probs, 5)
+    return {class_names[idx]: float(prob) for idx, prob in zip(top5_idx, top5_probs)}
+
+# ===============================
+# 📌 Gradio Interface
+# ===============================
+demo = gr.Interface(
+    fn=predict,
+    inputs=gr.Image(type="pil"),
+    outputs=gr.Label(num_top_classes=5),
+    title="🚗 Car Brand Classifier",
+    description="Upload a car image and the model predicts the brand (Top-5)."
+)
+
+if __name__ == "__main__":
+    demo.launch()

best_model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c15273cfeec798e0051940a97c1c18cf51a14a995e8aa155a9d98eeade1cc043
+size 98650771

requirements.txt

@@ -0,0 +1,4 @@
+torch
+torchvision
+gradio
+Pillow