add a model

laion/CLIP-ViT-H-14-laion2B-s32B-b79K

app.py

@@ -6,6 +6,10 @@ model = CLIPModel.from_pretrained(clip_path).eval()
 processor = AutoProcessor.from_pretrained(clip_path)
 
 
+clip_path2 = "laion/CLIP-ViT-H-14-laion2B-s32B-b79K"
+model2 = CLIPModel.from_pretrained(clip_path2).eval()
+processor2 = AutoProcessor.from_pretrained(clip_path2)
+
 async def predict(init_image, labels_level1):
     if init_image is None:
         return "", ""
@@ -16,13 +20,19 @@ async def predict(init_image, labels_level1):
     inputs = processor(
         text=split_labels, images=init_image, return_tensors="pt", padding=True
     )
-
+    
+    inputs2 = processor2(
+        text=split_labels, images=init_image, return_tensors="pt", padding=True
+    )
     outputs = model(**inputs)
+    outputs2 = model2(**inputs2)
     logits_per_image = outputs.logits_per_image  # this is the image-text similarity score
 
+    logits_per_image2 = outputs2.logits_per_image  # this is the image-text similarity score
 
     for i in range(len(split_labels)):
-        ret_str += split_labels[i] + ": " + str(logits_per_image[0][i]) + "\n"
+        ret_str += split_labels[i] + ": " + str(float(logits_per_image[0][i]))
+                + ", " + str(float(logits_per_image2[0][i])) + "\n"
 
     return ret_str, ret_str