qq

app.py

@@ -16,7 +16,7 @@ def predict(image):
     image = converter.enhance(0.5)
     image = image.convert("L")
     image = preprocess(image).unsqueeze(0).to(device)
-    text = clip.tokenize([f"a character of origin {c}" for c in labels]).to(device)
+    text = clip.tokenize([f"a character of origin: {c}" for c in labels]).to(device)
 
     with torch.inference_mode():
         logits_per_image, logits_per_text = model(image, text)