def initialize()

in templates/inference-endpoints/preprocessing/1/model.py [0:0]

• 19 lines of code
• 4 McCabe index (conditional complexity)

    def initialize(self, args):
        """`initialize` is called only once when the model is being loaded.
        Implementing `initialize` function is optional. This function allows
        the model to initialize any state associated with this model.
        Parameters
        ----------
        args : dict
          Both keys and values are strings. The dictionary keys and values are:
          * model_config: A JSON string containing the model configuration
          * model_instance_kind: A string containing model instance kind
          * model_instance_device_id: A string containing model instance device ID
          * model_repository: Model repository path
          * model_version: Model version
          * model_name: Model name
        """
        # Parse model configs
        model_config = json.loads(args['model_config'])
        tokenizer_dir = Path(model_config['parameters']['tokenizer_dir']['string_value'])
        tokenizer_path = tokenizer_dir.joinpath("tokenizer.json")
        pad_to_multiple_of = int(model_config['parameters']['pad_to_multiple_of']['string_value'])

        special_tokens_map_path = tokenizer_dir.joinpath("special_tokens_map.json")
        with open(special_tokens_map_path, "r", encoding="utf-8") as special_tokens_f:
            special_tokens_map = json.load(special_tokens_f)

        self.tokenizer = Tokenizer.from_file(str(tokenizer_path))

        if "eos_token" in special_tokens_map:
            eos_token = special_tokens_map["eos_token"]["content"]
            eos_token_id = self.tokenizer.encode(eos_token, add_special_tokens=False).ids[0]

            # self.tokenizer.enable_padding(
            #     direction="left", pad_id=eos_token_id, pad_token=eos_token, pad_to_multiple_of=pad_to_multiple_of
            # )
            self.pad_token = eos_token
            self.pad_token_id = eos_token_id

        # Parse model output configs and convert Triton types to numpy types
        for name in INPUT_NAMES:
            dtype = pb_utils.triton_string_to_numpy(
                pb_utils.get_output_config_by_name(model_config, name)['data_type']
            )

            setattr(self, name.lower() + "_dtype", dtype)