2noise/ChatTTS · error · ValueError

Unknown tokenizer mode: {self.tokenizer_mode}. Must be eithe

Error message

Unknown tokenizer mode: {self.tokenizer_mode}. Must be either 'auto' or 'slow'.

What it means

Error "Unknown tokenizer mode: {self.tokenizer_mode}. Must be either 'auto' or 'slow'." thrown in 2noise/ChatTTS.

Source

Thrown at ChatTTS/model/velocity/configs.py:153

            raise ValueError(
                f"load format '{load_format}' is not supported in ROCm. "
                f"Supported load format are "
                f"{rocm_supported_load_format}"
            )

        # TODO: Remove this check once HF updates the pt weights of Mixtral.
        architectures = getattr(self.hf_config, "architectures", [])
        if "MixtralForCausalLM" in architectures and load_format == "pt":
            raise ValueError(
                "Currently, the 'pt' format is not supported for Mixtral. "
                "Please use the 'safetensors' format instead. "
            )
        self.load_format = load_format

    def _verify_tokenizer_mode(self) -> None:
        tokenizer_mode = self.tokenizer_mode.lower()
        if tokenizer_mode not in ["auto", "slow"]:
            raise ValueError(
                f"Unknown tokenizer mode: {self.tokenizer_mode}. Must be "
                "either 'auto' or 'slow'."
            )
        self.tokenizer_mode = tokenizer_mode

    def _verify_quantization(self) -> None:
        supported_quantization = ["awq", "gptq", "squeezellm"]
        rocm_not_supported_quantization = ["awq"]
        if self.quantization is not None:
            self.quantization = self.quantization.lower()

        # Parse quantization method from the HF model config, if available.
        hf_quant_config = getattr(self.hf_config, "quantization_config", None)
        if hf_quant_config is not None:
            hf_quant_method = str(hf_quant_config["quant_method"]).lower()
            if self.quantization is None:
                self.quantization = hf_quant_method
            elif self.quantization != hf_quant_method:

View on GitHub (pinned to 77b89ee281)

When it happens

Trigger: Thrown at ChatTTS/model/velocity/configs.py:153 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of 2noise/ChatTTS@77b89ee281 (2026-08-26). Data as JSON: /api/errors/a2442b8f069473e7. Report an issue: GitHub.