Update handler.py

handler.py

@@ -1,5 +1,4 @@
 from typing import Dict, Any
-import torch
 from transformers import AutoProcessor, Qwen2VLForConditionalGeneration
 from PIL import Image
 import io
@@ -8,21 +7,16 @@ import requests
 
 class EndpointHandler():
     def __init__(self, path=""):
-        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-        self.model = Qwen2VLForConditionalGeneration.from_pretrained(path).to(self.device)
         self.processor = AutoProcessor.from_pretrained(path)
+        self.model = Qwen2VLForConditionalGeneration.from_pretrained(path)
 
     def __call__(self, data: Any) -> Dict[str, Any]:
-        default_prompt = "Describe this image."
+
+        image_input = data.get('image', None)
+        text_input = data.get('text', None)
         
-        if isinstance(data, (bytes, bytearray)):
-            image = Image.open(io.BytesIO(data)).convert('RGB')
-            text_input = default_prompt
-        elif isinstance(data, dict):
-            image_input = data.get('image', None)
-            text_input = data.get('text', default_prompt)
-            if image_input is None:
-                return {"error": "No image provided."}
+
+        if isinstance(data, dict):
             if image_input.startswith('http'):
                 image = Image.open(requests.get(image_input, stream=True).raw).convert('RGB')
             else: