initial

app.py

@@ -0,0 +1,57 @@
+import torch
+import gradio as gr
+from PIL import Image
+import torchvision.transforms.functional as TF
+import torchvision.models as models
+import torch.nn as nn
+
+# Load pre-trained EfficientNet
+model = models.efficientnet_b4()
+num_features = model.classifier[1].in_features
+model.classifier[1] = nn.Linear(num_features, 2)
+
+model.load_state_dict(torch.load("model/deep-image-squish-predictor-V0.pth", map_location=torch.device('cpu')))
+model.eval()
+
+def predict(image):
+    width, height = image.size
+    ratio = width / height
+    if(width > height):
+      height = int(256 * ratio)
+      width = 256
+    else:
+      width = int(256 / ratio)
+      height = 256
+    
+    image = TF.resize(image, (height, width))
+
+    padded_image = Image.new("RGB", (256, 256))
+    padded_image.paste(image, (0, 0))
+    image_tensor = TF.to_tensor(padded_image).unsqueeze(0)
+
+    # Predict the squish ratio
+    with torch.no_grad():
+        output = model(image_tensor)
+    
+    wsr, hsr = output.squeeze().tolist()
+    return f"Squish Ratio: (Width, Height)= ({wsr:.2f}, {hsr:.2f})"
+
+# Define the examples (provide paths to example images)
+examples = [
+    ["example_images/image1.jpg"],
+    ["example_images/image2.jpg"],
+    ["example_images/image3.jpg"]
+]
+
+# Create the Gradio interface
+iface = gr.Interface(
+    fn=predict,
+    inputs=gr.Image(type="pil"),
+    outputs="text",
+    examples=examples,
+    title="Deep Image Squish Predictor",
+    description="Upload an image to see the predicted squish ratios."
+)
+
+# Launch the interface
+iface.launch()

model/deep-image-squish-predictor-V0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:914ddca914511e0a085c5cf2bc1cd205fc37e56d88d6b9099c850c60515746b1
+size 70927906