Debugging handler.py

handler.py

@@ -1,4 +1,4 @@
-from typing import Dict
+from typing import Dict, Any
 import requests
 from transformers import CLIPProcessor, CLIPModel
 from PIL import Image
@@ -10,16 +10,18 @@ class EndpointHandler:
         self.processor = CLIPProcessor.from_pretrained(path)
         self.model = CLIPModel.from_pretrained(path)
 
-    def __call__(self, data: Dict) -> Dict:
-        text = data.pop("text")
+    def __call__(self, data: Dict[str, Any]) -> Dict:
+        print("this shows the custom endpoint handler is being called")
+        inputs = data.pop("inputs", data)
+        text = inputs.pop("text")
         if "image_url" in data:
-            image_url = data.pop("image_url")
+            image_url = inputs.pop("image_url")
             image = Image.open(requests.get(image_url, stream=True).raw)
         else:
-            image = data.pop("image")
-        inputs = self.processor(text=text, images=image,
-                                return_tensors="pt", padding=True, truncation=True)
-        outputs = self.model(**inputs)
+            image = inputs.pop("image")
+        processed_inputs = self.processor(text=text, images=image,
+                                          return_tensors="pt", padding=True, truncation=True)
+        outputs = self.model(**processed_inputs)
         embedding_similarity = cosine_similarity(outputs.text_embeds.detach().numpy(),
                                                  outputs.image_embeds.detach().numpy())[0][0].item()
         return {"text_embedding": outputs.text_embeds[0].tolist(),