chore: remove zerogpu

app.py

@@ -1,83 +1,26 @@
-import os
-# Set this before importing torch to avoid issues
-os.environ["GRADIO_ALLOW_FLAGGING"] = "never"
-
 import torch
 import gradio as gr
 import requests
-from torchvision import transforms, models
+from torchvision import transforms
+
+# Load model
+model = torch.hub.load('pytorch/vision:v0.6.0', 'resnet18', pretrained=True).eval()
 
 # Download human-readable labels for ImageNet.
 response = requests.get("https://git.io/JJkYN")
 labels = response.text.split("\n")
 
-# Load ResNet-34 model with pre-trained weights
-model = models.resnet34(weights=models.ResNet34_Weights.IMAGENET1K_V1)
-model.eval()
-
-# Preprocessing pipeline
-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]),
-])
-
-def predict_without_gpu(inp):
-    """CPU-only prediction for fallback"""
-    if inp is None:
-        return {}
-    
-    input_tensor = preprocess(inp).unsqueeze(0)
-    
+def predict(inp):
+    inp = transforms.ToTensor()(inp).unsqueeze(0)
     with torch.no_grad():
-        output = model(input_tensor)
-        prediction = torch.nn.functional.softmax(output[0], dim=0)
-        confidences = {labels[i]: float(prediction[i]) for i in range(1000)}
-    
+        prediction = torch.nn.functional.softmax(model(inp)[0], dim=0)
+        confidences = {labels[i]: float(prediction[i]) for i in range(1000)}    
     return confidences
 
-# Try to use GPU if available
-try:
-    import spaces
-    
-    @spaces.GPU(duration=60)
-    def predict(inp):
-        """GPU-accelerated prediction"""
-        if inp is None:
-            return {}
-        
-        device = torch.device("cuda")
-        model.to(device)
-        
-        input_tensor = preprocess(inp).unsqueeze(0).to(device)
-        
-        with torch.no_grad():
-            output = model(input_tensor)
-            prediction = torch.nn.functional.softmax(output[0], dim=0)
-            prediction = prediction.cpu()
-            confidences = {labels[i]: float(prediction[i]) for i in range(1000)}
-        
-        return confidences
-    
-    print("Using GPU-accelerated prediction")
-    prediction_fn = predict
-    
-except Exception as e:
-    print(f"GPU initialization failed: {e}")
-    print("Using CPU-only prediction")
-    prediction_fn = predict_without_gpu
-
-# Create the Gradio interface
-demo = gr.Interface(
-    fn=prediction_fn,
+gr.Interface(
+    fn=predict, 
     inputs=gr.Image(type="pil"),
     outputs=gr.Label(num_top_classes=3),
-    title="Image Classifier",
-    description="Upload an image to classify it using ResNet-34",
+    examples=["lion.jpg", "cheetah.jpg", "cat.avif", "hot-dog.avif", "llama.jpg", "medieval_knight.jpg"],
     css=".footer{display:none !important}"
-)
-
-# Launch without problematic parameters
-if __name__ == "__main__":
-    demo.launch()
+).launch()

requirements.txt

@@ -1,6 +1,5 @@
 torch
 gradio
-spaces
 numpy
 torchvision
-Pillow
+Pillow