handler.py

from transformers import AutoModelForCausalLM, AutoTokenizer
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"
        )
        self.model.eval()

    def __call__(self, data):
        inputs = data.pop("inputs", "")
        parameters = data.pop("parameters", {})

        max_new_tokens = parameters.get("max_new_tokens", 128)
        temperature = parameters.get("temperature", 0.7)
        top_p = parameters.get("top_p", 0.9)
        repetition_penalty = parameters.get("repetition_penalty", 1.1)

        tokenized = self.tokenizer(inputs, return_tensors="pt").to(self.model.device)

        with torch.no_grad():
            outputs = self.model.generate(
                **tokenized,
                max_new_tokens=max_new_tokens,
                temperature=temperature,
                top_p=top_p,
                repetition_penalty=repetition_penalty,
                do_sample=temperature > 0,
            )

        # Decode only the NEW tokens (exclude the input prompt)
        new_tokens = outputs[0][tokenized["input_ids"].shape[1]:]
        result = self.tokenizer.decode(new_tokens, skip_special_tokens=True)

        return [{"generated_text": result}]