Added model and application file

FaKe-ViT-B16.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a2d9f5edce776c627c3797b1f1a6be5d243a188ce39b9546da2ee031b363c30
+size 343286022

app.py

@@ -0,0 +1,59 @@
+import gradio as gr
+from PIL import Image
+import torch
+import torchvision.models as models
+from torchvision.transforms import v2 as transforms
+import os
+
+class_names = ['AI-Generated Image', "Real/Non-AI-Generated Image"]
+
+# Downloading the model
+# model = models.vit_b_16()
+weights_path = "FaKe-ViT-B16.pth"
+model = torch.load(weights_path).to("cpu")
+model.eval()
+# Preprocessing the image
+preprocess = transforms.Compose([
+    transforms.Resize(256),
+    transforms.CenterCrop(224),
+    transforms.ToTensor(),
+    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+])
+
+# Define the prediction function
+def predict_image(image):
+    # inp = Image.fromarray(inp.astype('uint8'), 'RGB')
+    image = preprocess(image)
+    if image.shape[0] != 3:
+        image = image[:3, :, :]
+    image = image.unsqueeze(0)
+    with torch.inference_mode():
+        output = model(image)
+    output1 = torch.argmax(torch.softmax(output,dim=1),dim=1).item()
+    return class_names[output1]
+
+# def image_mod(image):
+#     return image.rotate(45)
+
+
+demo = gr.Interface(
+    predict_image,
+    gr.Image(image_mode="RGB",type="pil"),
+    "text",
+    flagging_options=["incorrect prediction"],
+    # examples=[
+    #     os.path.join(os.path.dirname(__file__), "images/cheetah1.jpg"),
+    #     os.path.join(os.path.dirname(__file__), "images/lion.jpg"),
+    #     os.path.join(os.path.dirname(__file__), "images/logo.png"),
+    #     os.path.join(os.path.dirname(__file__), "images/tower.jpg"),
+    # ],
+    title="FaKe-ViT-B/16: AI-Generated Image Detection using Vision Transformer(ViT-B/16)",
+    description="This is a demo to detect AI-Generated images using Vision Transformer(ViT-B/16). Upload an image and the model will predict whether the image is AI-Generated or Real",
+    css=""".gr-header, .gr-text {
+        font-size: 20px;
+    }""",
+    article=" \nBased on the paper:'An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale', Alexey et al.\nDataset: 'Fake or Real competition dataset' at https://huggingface.co/datasets/mncai/Fake_or_Real_Competition_Dataset"
+)
+
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,5 @@
+gradio
+uvicorn
+starlette
+torch
+torchvision