handler.py

from typing import Dict, List, Any
from transformers import pipeline, AutoTokenizer, BartForConditionalGeneration

class EndpointHandler():
    def __init__(self, path=""):
        self.device = 'cuda' if torch.cuda.is_available() else 'cpu'
        try:
            self.model = BartForConditionalGeneration.from_pretrained(path).to(self.device)
            self.tokenizer = AutoTokenizer.from_pretrained(path)
        except Exception as e:
            print(f"Error loading model or tokenizer from path {path}: {e}")
        

    def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]:
        """
       data args:
            inputs (:obj: `str`)
            date (:obj: `str`)
      Return:
            A :obj:`list` | `dict`: will be serialized and returned
        """
        # get inputs
  

        inputs = data.get("inputs", "")
        if not inputs:
            return [{"error": "No inputs provided"}]

        tokenized_input = self.tokenizer(inputs, return_tensors="pt", truncation=True, max_length=1024, padding="max_length")
        tokenized_input = tokenized_input.to(self.device)  # Move input tensors to the same device as model

        summary_ids = self.model.generate(**tokenized_input, max_length=256, do_sample=True, top_p=0.8)

        summary_text = self.tokenizer.decode(summary_ids[0], skip_special_tokens=True)

        return [{"summary": summary_text}]