invoke-ai/InvokeAI · error · ValueError

Source longer side ( px) is smaller than the Wan pixel grid…

Error message

Source longer side ({long_side}px) is smaller than the Wan pixel grid ({multiple}px). Use an input image at least {multiple}px on its longer side.

What it means

Wan models require output dimensions snapped to a pixel grid (16px for 8x-VAE models, 32px for the 16x-VAE TI2V-5B). If the source's longer side is smaller than that grid multiple, the clamp would silently collapse the result to multiple x multiple, destroying the requested aspect ratio. The node raises instead so the workflow author fixes the input image.

Solutions

Upscale or replace the input image so its longer side is at least the grid multiple (>=16px, or >=32px for TI2V-5B)
Use the ideal-dimensions node matching your model's multiple (16 vs 32)
Check upstream resize/crop nodes that may have shrunk the image

Example fix

// before
inputImage: 12x20px, multiple=16  -> error
// after
inputImage: 480x800px (or upscaled) then Wan Ideal Dimensions

Defensive patterns

Strategy: validation

Validate before calling

multiple = 32 if model_is_16x_vae else 16
if max(width, height) < multiple:
    raise ValueError(f"image longer side must be >= {multiple}px")

Try / catch

try:
    dims = ideal_dims.invoke(context)
except ValueError as e:
    if 'smaller than the Wan pixel grid' in str(e):
        upscale_image_to_at_least(multiple)
    else:
        raise

Prevention

Use input images with a reasonable resolution (>= grid multiple)
Match the multiple (16 vs 32) to your Wan model variant
Avoid aggressive downsizing upstream in the pipeline

When it happens

Trigger: Feeding a tiny image (longer side < 16 or < 32 px, e.g. a 12x12 thumbnail) into Wan Ideal Dimensions; using multiple=32 (TI2V-5B) with a small image whose long side is under 32px.

Common situations: Downscaled or placeholder images in automated pipelines; switching to the 16x-VAE model (multiple=32) without enlarging previously-fine small inputs.

Related errors

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/0c879e9663797ef7. Report an issue: GitHub.

Appendix: source

Thrown at invokeai/app/invocations/wan_ideal_dimensions.py:70

    multiple: int,
) -> tuple[int, int]:
    """Scale a source W×H so its shorter side equals ``target_short_side``, then
    snap each dimension to ``multiple`` using the requested rounding mode.

    ``multiple`` is the Wan pixel-grid constraint (16 for the 8x-VAE I2V/T2V
    models, 32 for the 16x-VAE TI2V-5B). Shared by both ideal-dimensions nodes.
    """
    short = min(width, height)
    if short <= 0:
        raise ValueError("Source dimensions must be positive.")

    # Reject sources so narrow that the scaled long side is still under one Wan
    # pixel grid. The downstream clamp to ``max(w, multiple)`` would otherwise
    # silently return multiple×multiple, which has no relation to the requested
    # aspect ratio — better to fail fast and have the workflow author fix inputs.
    long_side = max(width, height)
    if long_side < multiple:
        raise ValueError(
            f"Source longer side ({long_side}px) is smaller than the Wan pixel grid ({multiple}px). "
            f"Use an input image at least {multiple}px on its longer side."
        )

    scale = target_short_side / short
    raw_w = width * scale
    raw_h = height * scale

    if rounding == "floor":
        w = int(raw_w // multiple) * multiple
        h = int(raw_h // multiple) * multiple
    elif rounding == "ceiling":
        w = int(math.ceil(raw_w / multiple)) * multiple
        h = int(math.ceil(raw_h / multiple)) * multiple
    else:  # nearest
        w = round(raw_w / multiple) * multiple
        h = round(raw_h / multiple) * multiple

View on GitHub (pinned to 0b6a024f2f)