Update handler.py

handler.py

@@ -3,9 +3,13 @@ from typing import  Dict, List, Any
 from PIL import Image
 import torch
 import os
+import io
+import base64
 from io import BytesIO
 # from transformers import BlipForConditionalGeneration, BlipProcessor
 from transformers import Blip2Processor, Blip2ForConditionalGeneration
+
+
 # -
 
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
@@ -30,19 +34,31 @@ class EndpointHandler():
             A :obj:`dict`:. The object returned should be a dict of one list like {"captions": ["A hugging face at the office"]} containing :
                 - "caption": A string corresponding to the generated caption.
         """
+        print(data)
         inputs = data.pop("inputs", data)
         parameters = data.pop("parameters", {})
+        print(input)
+        image_bytes = base64.b64decode(inputs)
+        image_io = io.BytesIO(image_bytes)
+        image = Image.open(image_io)
+
+        inputs = processor(images=image, return_tensors="pt").to(device, torch.float16)
+        pixel_values = inputs.pixel_values
+        
+        generated_ids = model.generate(pixel_values=pixel_values, max_length=25)
+        generated_caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+        print(generated_caption)
  
-        raw_images = [Image.open(BytesIO(_img)) for _img in inputs]
+        # raw_images = [Image.open(BytesIO(_img)) for _img in inputs]
                                      
-        processed_image = self.processor(images=raw_images, return_tensors="pt") 
-        processed_image["pixel_values"] = processed_image["pixel_values"].to(device)
-        processed_image = {**processed_image, **parameters}
+        # processed_image = self.processor(images=raw_images, return_tensors="pt") 
+        # processed_image["pixel_values"] = processed_image["pixel_values"].to(device)
+        # processed_image = {**processed_image, **parameters}
         
-        with torch.no_grad():
-            out = self.model.generate(
-                **processed_image
-            )
-        captions = self.processor.batch_decode(out, skip_special_tokens=True)
+        # with torch.no_grad():
+        #     out = self.model.generate(
+        #         **processed_image
+        #     )
+        # captions = self.processor.batch_decode(out, skip_special_tokens=True)
 
         return {"captions": captions}