Works, and examples

app.py

@@ -13,26 +13,31 @@ print("Loaded model.")
 def process(image, prompt):
     print("Inferring...")
     image = preprocess(image).unsqueeze(0).to(device)
-    print(image)
+    print("Image: ", image)
 
     prompts = prompt.split("\n")
+    print("Prompts: ", prompts)
     text = clip.tokenize(prompts).to(device)
-    print(text)
+    print("Tokens: ", text)
 
     with torch.no_grad():
         logits_per_image, logits_per_text = model(image, text)
-        probs = logits_per_image.softmax(dim=-1).cpu().numpy()
-        print(probs)
+        probs = logits_per_image.softmax(dim=-1).cpu()
+        print("Probs: ", probs)
 
-        return dict(zip(prompts, probs))
+        return {k: v.item() for (k,v) in zip(prompts, probs[0])}
 
 
 iface = gr.Interface(
     fn=process,
     inputs=[
-        gr.Image(type="pil"),
+        gr.Image(type="pil", label="Image"),
         gr.Textbox(lines=5, label="Prompts (newline-separated)"),
     ],
     outputs="label",
+    examples=[
+        ["dog.jpg", "a photo of a dog\na photo of a cat"],
+        ["cat.jpg", "a photo of a dog\na photo of a cat"],
+    ]
 )
 iface.launch()