Update app.py

app.py

@@ -5,10 +5,11 @@ import torch
 from PIL import Image
 import base64
 from io import BytesIO
+from decimal import Decimal  # Import the Decimal module
 
 # Load the CLIP model
 model, preprocess = clip.load("ViT-B/32")
-device = "cuda" if torch.cuda.is_available() else "cpu"
+device = "cuda" if torch.cuda.isavailable() else "cpu"
 model.to(device).eval()
 
 # Define a function to find similarity
@@ -17,7 +18,7 @@ def find_similarity(base64_image, text_input):
     image_bytes = base64.b64decode(base64_image)
 
     # Convert the bytes to a PIL image
-    image = Image.open(BytesIO(image_bytes))
+    image = Image.open(BytesIO(image_bytes)
 
     # Preprocess the image
     image = preprocess(image).unsqueeze(0).to(device)
@@ -33,7 +34,10 @@ def find_similarity(base64_image, text_input):
     # Calculate cosine similarity
     similarity = (image_features @ text_features.T).squeeze(0).cpu().numpy()
 
-    return similarity[0,0]
+    # Convert the similarity score to a Decimal
+    similarity_decimal = Decimal(similarity)
+
+    return similarity_decimal  # Return the similarity score as a Decimal
 
 # Create a Gradio interface
 iface = gr.Interface(
@@ -42,7 +46,7 @@ iface = gr.Interface(
         gr.inputs.Textbox(label="Base64 Image", lines=8),
         "text"
     ],
-    outputs="number",
+    outputs="text",  # Set the output type to "text"
     live=True,
     interpretation="default",
     title="CLIP Model Image-Text Cosine Similarity",