handler.py

# handler.py
from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
import torch

class EndpointHandler:
    def __init__(self, path=""):
        self.tokenizer = AutoTokenizer.from_pretrained(path)
        self.model = AutoModelForCausalLM.from_pretrained(
            path,
            torch_dtype=torch.float16,
            device_map="auto",
            load_in_4bit=True,
        )
        self.pipeline = pipeline(
            "text-generation",
            model=self.model,
            tokenizer=self.tokenizer,
        )

    def __call__(self, data):
        messages = data.get("inputs", {}).get("messages", [])
        prompt = self.tokenizer.apply_chat_template(
            messages, tokenize=False, add_generation_prompt=True
        )
        result = self.pipeline(
            prompt,
            max_new_tokens=data.get("parameters", {}).get("max_tokens", 500),
            temperature=data.get("parameters", {}).get("temperature", 0.45),
            do_sample=True,
        )
        text = result[0]["generated_text"][len(prompt):]
        return {"choices": [{"message": {"role": "assistant", "content": text}}]}