invoke-ai/InvokeAI · error · ValueError

The requested denoising range does not contain any…

Error message

The requested denoising range does not contain any effective denoising steps at the configured step count. Increase denoising_end, decrease denoising_start, or increase steps.

What it means

`_validate_effective_schedule` checks that the computed denoising step window [start_idx, end_idx) contains at least one step. When the start/end fractions are so close together (or steps so few) that both indices round to the same value, no effective denoising occurs and this ValueError is raised.

Solutions

1. Increase denoising_end so the range spans at least one timestep at the current step count.
2. Decrease denoising_start to widen the effective window.
3. Increase `steps` so the fractional range maps to at least one discrete step.

Example fix

// before: range rounds to zero steps at steps=10
denoising_start=0.50; denoising_end=0.505; steps=10
// after: widen the range or add steps
denoising_start=0.50; denoising_end=0.60; steps=10

def effective_steps(start: float, end: float, steps: int) -> int:
    s = int(round(start * (steps - 1)))
    e = int(round(end * (steps - 1)))
    return max(0, e - s)
# guard: if effective_steps(start, end, steps) < 1: widen range or raise steps

def denoise_range_yields_steps(start: float, end: float, steps: int) -> bool:
    return int(round(end * (steps - 1))) > int(round(start * (steps - 1)))

try:
    out = invoke_krea2_denoise(denoising_start=start, denoising_end=end, steps=steps)
except ValueError as e:
    if "does not contain any effective denoising steps" in str(e):
        end = min(1.0, start + max(1 / steps, 0.1))
        out = invoke_krea2_denoise(denoising_start=start, denoising_end=end, steps=steps)
    else:
        raise

When it happens

Trigger: denoising_start and denoising_end very close together (e.g. 0.50 to 0.505) with a small steps count, so both map to the same timestep index; end_idx computed <= start_idx after rounding.

Common situations: Img2img/refiner workflows with a narrow denoise window; users lowering steps to speed up runs after setting a fine-grained start/end range; UI sliders producing nearly identical values.

Related errors

• denoising_start must be less than denoising_end.
• denoising_start should be 0 when initial latents are not…
• At least one Krea-2 conditioning is required.
• cfg_scale list has values but the model is configured for…
• cfg_scale values must be finite.

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/4a92d38bfffb5a95. Report an issue: GitHub.

Appendix: source

Thrown at invokeai/app/invocations/krea2_denoise.py:224

        if isinstance(self.cfg_scale, float):
            return [self.cfg_scale] * num_timesteps
        if isinstance(self.cfg_scale, list):
            if len(self.cfg_scale) != num_timesteps:
                raise ValueError(
                    f"cfg_scale list has {len(self.cfg_scale)} values but the model is configured for "
                    f"{num_timesteps} steps. Provide one CFG value per configured step (or a single float)."
                )
            return self.cfg_scale
        raise ValueError(f"Invalid CFG scale type: {type(self.cfg_scale)}")

    @staticmethod
    def _should_apply_cfg_for_step(cfg_scale: float, *, has_negative_conditioning: bool) -> bool:
        return has_negative_conditioning and cfg_scale > 1.0

    @staticmethod
    def _validate_effective_schedule(*, start_idx: int, end_idx: int) -> None:
        if end_idx <= start_idx:
            raise ValueError(
                "The requested denoising range does not contain any effective denoising steps at the configured "
                "step count. Increase denoising_end, decrease denoising_start, or increase steps."
            )

    def _validate_inputs(self) -> None:
        if self.denoising_start >= self.denoising_end:
            raise ValueError("denoising_start must be less than denoising_end.")
        if self.denoise_mask is not None and self.latents is None:
            raise ValueError("Initial latents are required when a denoise mask is provided.")

    def _is_distilled(self, context: InvocationContext) -> bool:
        """Whether the transformer is the distilled Turbo checkpoint (fixed mu) vs. Raw (dynamic mu).

        Prefer the classified variant (works for diffusers, single-file and GGUF alike); fall back to
        the pipeline-level ``is_distilled`` flag in model_index.json, then default to distilled.

        A failed config lookup is a real error and is allowed to propagate — silently defaulting to the
        Turbo shift would apply the wrong sampling schedule to a Raw model.

View on GitHub (pinned to 0b6a024f2f)