Add application file

app.py

@@ -26,15 +26,25 @@ model.eval()
 
 
 def predict(im):
-    if isinstance(im, dict):
+    if isinstance(im, dict):  # For sketchpad input
         im = im['composite']
-    x = torch.tensor(im, dtype=torch.float32).unsqueeze(0).unsqueeze(0) / 255.0
+    
+    # Convert to grayscale and resize to 28x28
+    import cv2
+    im_gray = cv2.cvtColor(im, cv2.COLOR_RGB2GRAY)
+    im_resized = cv2.resize(im_gray, (28, 28))
+
+    # Convert to tensor and normalize
+    x = torch.tensor(im_resized, dtype=torch.float32).unsqueeze(0).unsqueeze(0) / 255.0
+
     with torch.no_grad():
         out = model(x)
+
     probabilities = torch.nn.functional.softmax(out[0], dim=0)
     values, indices = torch.topk(probabilities, 5)
     return {LABELS[i]: v.item() for i, v in zip(indices, values)}
 
+
 interface = gr.Interface(
     predict,
     inputs="sketchpad",

requirements.txt

@@ -1,3 +1,4 @@
 torch
+opencv-python
 gradio
 pathlib