captions along with probs

app.py

@@ -10,12 +10,12 @@ def inference(input_img, captions):
     outputs = clip(**inputs)
     # this is the image-text similarity score
     logits_per_image = outputs.logits_per_image
-    probs = logits_per_image.softmax(dim=1)
-    confidences = {i: probs[i] for i in range(len(probs))}
+    probs = logits_per_image.softmax(dim=1).tolist()[0]
+    confidences = {captions_list[i][:30]: probs[i] for i in range(len(probs))}
     return confidences
 
 title = "CLIP Inference: Application using a pretrained CLIP model"
-description = "An application using Gradio interface that accepts an image and some captions, and displays a probability score with which each caption describes the image "
+description = "An application to predict the appropriate caption for an image"
 
 examples = [
     ["examples/woman_standing.jpg","woman standing inside a house, a photo of dog, running water, cupboard, home interiors"],