Upload 3 files

app.py

@@ -0,0 +1,53 @@
+import gradio as gr
+import torch
+from torchvision import models, transforms
+from PIL import Image
+import torch.nn.functional as F
+import torch.nn as nn
+
+# Class labels
+class_names = ['fake', 'real']
+
+# Image transform
+data_transforms = transforms.Compose([
+    transforms.Resize((224, 224)),
+    transforms.ToTensor(),
+    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
+])
+
+# Load model
+def load_model():
+    model = models.densenet121(weights='IMAGENET1K_V1')
+    model.classifier = nn.Sequential(
+        nn.Linear(1024, 512),
+        nn.ReLU(),
+        nn.Dropout(0.5),
+        nn.Linear(512, 2)
+    )
+    device = torch.device('cpu')  # Use CPU for Hugging Face
+    model = model.to(device)
+    checkpoint = torch.load("best_model.pth", map_location=device)
+    model.load_state_dict(checkpoint['model_state_dict'])
+    model.eval()
+    return model, device
+
+model, device = load_model()
+
+# Inference function
+def predict(image):
+    if image.mode == "RGBA":
+        image = image.convert("RGB")
+
+    image = data_transforms(image).unsqueeze(0).to(device)
+    with torch.no_grad():
+        outputs = model(image)
+        probs = F.softmax(outputs, dim=1)
+        conf, pred = torch.max(probs, 1)
+
+    label = class_names[pred.item()]
+    confidence = f"{conf.item() * 100:.2f}%"
+    return f"{label} ({confidence})"
+
+# Gradio interface
+demo = gr.Interface(fn=predict, inputs=gr.Image(type="pil"), outputs="text", title="Fake/Real Image Classifier")
+demo.launch()

best_model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b25e61fd341bbf0ab85e38bf838c9fe7939cea3d32565d3423ca9f70654eacf
+size 80199993

requirements.txt

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