handler.py

import torch
import transformers
from typing import Any, Dict
from transformers import AutoModelForCausalLM, AutoTokenizer


# class EndpointHandler():
#     def __init__(self, path=""):
#         model = AutoModelForCausalLM.from_pretrained(path,
#                                                      torch_dtype=torch.bfloat16,
#                                                      trust_remote_code=True,
#                                                      device_map="auto")
#         print(model.hf_device_map)
#         tokenizer = AutoTokenizer.from_pretrained(path)
#         #device = "cuda:0" if torch.cuda.is_available() else "cpu"
#         self.pipeline = transformers.pipeline('text-generation',
#                                               model=model,
#                                               tokenizer=tokenizer)

#     def __call__(self, data: Dict[str, Any]):
#         inputs = data.pop("inputs", data)
#         parameters = data.pop("parameters", {})
#         with torch.autocast(self.pipeline.device.type, dtype=torch.bfloat16):
#             outputs = self.pipeline(inputs,
#                                     **parameters)
#             return outputs


class EndpointHandler:
    def __init__(self, path=""):
        # load model and tokenizer from path
        self.tokenizer = AutoTokenizer.from_pretrained(path)
        self.device = "cuda" if torch.cuda.is_available() else "cpu"
        self.model = AutoModelForCausalLM.from_pretrained(path,
                                                          device_map="auto",
                                                          torch_dtype=torch.float16,
                                                          trust_remote_code=True)

    def __call__(self, data: Dict[str, Any]) -> Dict[str, str]:
        # process input
        inputs = data.pop("inputs", data)
        parameters = data.pop("parameters", {})
        return_full_text = parameters.pop("return_full_text", True)

        # preprocess
        inputs = self.tokenizer(inputs,
                                return_tensors="pt",
                                return_token_type_ids=False)
        inputs = inputs.to(self.device)
        input_len = len(inputs[0])

        outputs = self.model.generate(**inputs, **parameters)[0]

        if not return_full_text:
            outputs = outputs[input_len:]

        # postprocess the prediction
        prediction = self.tokenizer.decode(outputs,
                                           skip_special_tokens=True)

        return [{"generated_text": prediction}]