handler.py

from transformers import AutoProcessor, AutoModel

class EndpointHandler():
    def __init__(self, path=""):
        self.processor = AutoProcessor.from_pretrained(path)
        self.model = AutoModel.from_pretrained(path)

    def __call__(self, data): #-> List[Dict[str, Any]]

        # get inputs
        input_text = data['inputs']

        inputs = self.processor(
            text=input_text,
            return_tensors="pt",
            voice_preset = "v2/en_speaker_6"
            )
        speech_values = self.model.generate(**inputs, do_sample=True)
        sample_rate = self.model.generation_config.sample_rate


        return {'audio': speech_values.tolist(), 'sample_rate': sample_rate}