2noise/ChatTTS · error · ValueError

top_p must be in (0, 1], got {self.top_p}.

Error message

top_p must be in (0, 1], got {self.top_p}.

What it means

SamplingParams._verify_args enforces top_p in the range (0, 1] (open at zero). top_p is nucleus-sampling probability mass; 0 would keep no candidates and values above 1 are meaningless.

Source

Thrown at ChatTTS/model/velocity/sampling_params.py:206

        if not -2.0 <= self.presence_penalty <= 2.0:
            raise ValueError(
                "presence_penalty must be in [-2, 2], got " f"{self.presence_penalty}."
            )
        if not -2.0 <= self.frequency_penalty <= 2.0:
            raise ValueError(
                "frequency_penalty must be in [-2, 2], got "
                f"{self.frequency_penalty}."
            )
        if not 0.0 < self.repetition_penalty <= 2.0:
            raise ValueError(
                "repetition_penalty must be in (0, 2], got "
                f"{self.repetition_penalty}."
            )
        # if self.temperature < 0.0:
        #     raise ValueError(
        #         f"temperature must be non-negative, got {self.temperature}.")
        if not 0.0 < self.top_p <= 1.0:
            raise ValueError(f"top_p must be in (0, 1], got {self.top_p}.")
        if self.top_k < -1 or self.top_k == 0:
            raise ValueError(
                f"top_k must be -1 (disable), or at least 1, " f"got {self.top_k}."
            )
        if not 0.0 <= self.min_p <= 1.0:
            raise ValueError("min_p must be in [0, 1], got " f"{self.min_p}.")
        if self.max_tokens < 1:
            raise ValueError(f"max_tokens must be at least 1, got {self.max_tokens}.")
        if self.logprobs is not None and self.logprobs < 0:
            raise ValueError(f"logprobs must be non-negative, got {self.logprobs}.")
        if self.prompt_logprobs is not None and self.prompt_logprobs < 0:
            raise ValueError(
                f"prompt_logprobs must be non-negative, got " f"{self.prompt_logprobs}."
            )

    def _verify_beam_search(self) -> None:
        if self.best_of == 1:
            raise ValueError(

View on GitHub (pinned to 77b89ee281)

Solutions

1. For greedy/deterministic output use temperature=0 or top_k=1 rather than top_p=0
2. Keep top_p in (0, 1]; 1.0 disables nucleus filtering
3. Validate/clamp user input

Example fix

# before
params = SamplingParams(top_p=0)

# after
params = SamplingParams(temperature=0)  # greedy

def normalize_top_p(v: float) -> float:
    return min(max(float(v), 0.000001), 1.0)

def is_valid_top_p(v) -> bool:
    return isinstance(v, (int, float)) and 0.0 < v <= 1.0

When it happens

Trigger: Constructing SamplingParams with top_p=0, a negative value, or top_p > 1.0 (e.g. 1.5 passed through from user input).

Common situations: Trying to make sampling deterministic by setting top_p=0 (use temperature=0 / greedy instead); a typo like top_p=10; unvalidated API payloads.

Related errors

• n must be at least 1, got {self.n}.
• best_of must be greater than or equal to n, got n={self.n} a
• presence_penalty must be in [-2, 2], got {self.presence_pena
• frequency_penalty must be in [-2, 2], got {self.frequency_pe
• repetition_penalty must be in (0, 2], got {self.repetition_p

AI-assisted analysis of 2noise/ChatTTS@77b89ee281 (2026-08-26). Data as JSON: /api/errors/27ad59e5128833b9. Report an issue: GitHub.