Upload handler.py

handler.py

@@ -0,0 +1,23 @@
+from typing import Any, Dict
+from transformers import Blip2Processor, Blip2ForConditionalGeneration
+
+
+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.to(self.device)
+
+    def __call__(self, data: Dict[str, Any]) -> Dict[str, str]:
+        # process input
+        image = data.pop("image", data)
+        text = data.pop("text", data)
+
+        inputs = self.processor(images=image, text=prompt, return_tensors="pt").to(self.device, torch.float16)
+        generated_ids = self.model.generate(**inputs)
+        generated_text = self.processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
+
+        return [{"answer": generated_text}]