handler.py

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel, PeftConfig
import torch

class EndpointHandler:
    def __init__(self, path=""):
        config = PeftConfig.from_pretrained(path)
        self.tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)
        base_model = AutoModelForCausalLM.from_pretrained(
            config.base_model_name_or_path, 
            torch_dtype=torch.float16
        )
        self.model = PeftModel.from_pretrained(base_model, path)
        self.model.eval()

    def __call__(self, inputs):
        prompt = inputs.get("inputs", "")
        inputs = self.tokenizer(prompt, return_tensors="pt").to(self.model.device)
        with torch.no_grad():
            outputs = self.model.generate(**inputs, max_new_tokens=800)
        return {"generated_text": self.tokenizer.decode(outputs[0], skip_special_tokens=True)}