microsoft/VibeVoice · error · ValueError

missing`sample` as a required keyword argument

Error message

 missing`sample` as a required keyword argument

What it means

The second/third-order multistep update helper (same `timestep_list`-style signature as 57) raises this when `sample` is absent from both kwargs and args[2]. Note the message has drifted from the siblings: ` missing\`sample`` — a stray backtick and leading space in this vendored copy, so exact string matching on the message will not match the other two variants.

Source

Thrown at vibevoice/schedule/dpm_solver.py:846

        Args:
            model_output_list (`List[torch.Tensor]`):
                The direct outputs from learned diffusion model at current and latter timesteps.
            sample (`torch.Tensor`):
                A current instance of a sample created by diffusion process.

        Returns:
            `torch.Tensor`:
                The sample tensor at the previous timestep.
        """

        timestep_list = args[0] if len(args) > 0 else kwargs.pop("timestep_list", None)
        prev_timestep = args[1] if len(args) > 1 else kwargs.pop("prev_timestep", None)
        if sample is None:
            if len(args) > 2:
                sample = args[2]
            else:
                raise ValueError(" missing`sample` as a required keyword argument")
        if timestep_list is not None:
            deprecate(
                "timestep_list",
                "1.0.0",
                "Passing `timestep_list` is deprecated and has no effect as model output conversion is now handled via an internal counter `self.step_index`",
            )

        if prev_timestep is not None:
            deprecate(
                "prev_timestep",
                "1.0.0",
                "Passing `prev_timestep` is deprecated and has no effect as model output conversion is now handled via an internal counter `self.step_index`",
            )

        sigma_t, sigma_s0, sigma_s1, sigma_s2 = (
            self.sigmas[self.step_index + 1],
            self.sigmas[self.step_index],
            self.sigmas[self.step_index - 1],

View on GitHub (pinned to 94da20d98b)

Solutions

1. Pass the sample tensor: `...(model_output, sample=sample)`.
2. Use the public `step()` API instead of the internal update helpers.
3. When matching these errors by message, allow for formatting drift (match on 'sample' + 'required keyword' rather than the exact string).

Example fix

# before
prev = sched._multistep_solver_second_order_update_corrected(mo, t_list, t_prev)

# after
prev = sched._multistep_solver_second_order_update_corrected(mo, sample=sample)

assert sample is not None, "sample tensor is required"
prev = scheduler._multistep_solver_second_order_update_corrected(model_output, sample=sample)

import torch

def has_sample(sample) -> bool:
    return isinstance(sample, torch.Tensor)

try:
    prev = sched._multistep_solver_second_order_update_corrected(mo, sample=sample)
except ValueError as e:
    if "sample" in str(e) and "keyword argument" in str(e):
        raise TypeError("scheduler update helper called without sample") from e
    raise

When it happens

Trigger: Direct invocation like `scheduler._multistep_solver_second_order_update_corrected(model_output, timestep_list, prev_timestep)` without `sample=...`; subclass overrides dropping the argument.

Common situations: Custom high-order update experiments; adapting the scheduler for batched audio tokens; grep-based error handling that expects the sibling message format.

Related errors

• missing `sample` as a required keyword argument
• missing `sample` as a required keyword argument
• Must pass exactly one of `num_inference_steps` or `timesteps
• Can only pass one of `num_inference_steps` or `custom_timest
• Number of inference steps is 'None', you need to run 'set_ti

AI-assisted analysis of microsoft/VibeVoice@94da20d98b (2026-08-15). Data as JSON: /api/errors/51cfbaf60004551b. Report an issue: GitHub.