Update handler.py

handler.py

@@ -1,5 +1,4 @@
 from typing import Any, Dict
-import torch
 from transformers import Blip2Processor, Blip2ForConditionalGeneration
 
 
@@ -7,8 +6,8 @@ class EndpointHandler:
     def __init__(self, path=""):
         # load model and processor from path
         self.processor = Blip2Processor.from_pretrained(path)
-        self.model = Blip2ForConditionalGeneration.from_pretrained(path, torch_dtype=torch.float16)
-        self.device = "cuda" if torch.cuda.is_available() else "cpu"
+        self.model = Blip2ForConditionalGeneration.from_pretrained(path)
+        self.device = "cuda"
 
         self.model.to(self.device)
 
@@ -17,7 +16,7 @@ class EndpointHandler:
         image = data.pop("image", data)
         text = data.pop("text", data)
 
-        inputs = self.processor(images=image, text=text, return_tensors="pt").to(self.device, torch.float16)
+        inputs = self.processor(images=image, text=prompt, return_tensors="pt").to(self.device)
         generated_ids = self.model.generate(**inputs)
         generated_text = self.processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()