invoke-ai/InvokeAI · error · ValueError

Reference-image dimensions ({self.ref_image.width}x{self.ref

Error message

Reference-image dimensions ({self.ref_image.width}x{self.ref_image.height}) must match denoise dimensions ({self.width}x{self.height}).

What it means

The A14B reference condition tensor is spatially aligned with the denoise latents, so the reference image's width/height (baked in at encode time) must exactly match the denoise node's width/height. InvokeAI raises ValueError on any mismatch rather than resizing, which would corrupt the conditioning geometry.

Source

Thrown at invokeai/app/invocations/wan_video_denoise.py:213

            )

        # I2V condition tensor. Two flavours:
        #   * A14B I2V — [1, 20, T_lat, H_lat, W_lat] (4 mask + 16 latent channels).
        #     Concatenated to noise latents along the channel dim each step → 36ch.
        #   * TI2V-5B I2V — [1, 48, 1, H_lat, W_lat] (single latent frame, same
        #     channel count as the noise latents). Blended with noise via a
        #     first_frame_mask at every step (expand_timesteps path).
        # Variant dispatch happens via the condition tensor's channel count below.
        ref_condition: torch.Tensor | None = None
        if self.ref_image is not None:
            if variant not in (WanVariantType.I2V_A14B, WanVariantType.TI2V_5B):
                raise ValueError(
                    f"Reference-image conditioning is only supported by Wan 2.2 I2V variants "
                    f"(I2V-A14B or TI2V-5B). The selected transformer is {variant.value!r}. "
                    "Remove the Reference Image input or load an I2V variant."
                )
            if self.ref_image.width != self.width or self.ref_image.height != self.height:
                raise ValueError(
                    f"Reference-image dimensions ({self.ref_image.width}x{self.ref_image.height}) must "
                    f"match denoise dimensions ({self.width}x{self.height})."
                )
            # A14B encodes one condition tensor per pixel-frame count, so the
            # encoder's num_frames must match. TI2V-5B's condition is always
            # single-frame regardless of the output length, so the field's
            # num_frames is informational only and we skip this check.
            if variant == WanVariantType.I2V_A14B and self.ref_image.num_frames != self.num_frames:
                raise ValueError(
                    f"Reference-image num_frames ({self.ref_image.num_frames}) must match denoise "
                    f"num_frames ({self.num_frames}). Re-run the Reference Image - Wan 2.2 node with "
                    f"num_frames={self.num_frames}."
                )
            ref_condition = context.tensors.load(self.ref_image.condition_tensor_name).to(
                device=device, dtype=inference_dtype
            )
            _validate_ref_condition_shape(
                ref_condition,

View on GitHub (pinned to 0b6a024f2f)

Solutions

Set the Reference Image node's width/height to exactly the denoise node's width/height (or vice versa).
Use the same shared resolution constants/fields in both nodes so they cannot diverge.
Re-run the Reference Image - Wan 2.2 node after any resolution change.

Example fix

// before
ref_encoder.width, ref_encoder.height = 832, 480
denoise.width, denoise.height = 512, 512
// after
ref_encoder.width, ref_encoder.height = 512, 512
denoise.width, denoise.height = 512, 512

Defensive patterns

Strategy: validation

Validate before calling

if ref_image is not None and (ref_image.width != width or ref_image.height != height):
    raise ValueError(f"Ref {ref_image.width}x{ref_image.height} != denoise {width}x{height}; re-encode")

Try / catch

try:
    tensor = denoise._run_diffusion(context)
except ValueError as e:
    if "must match denoise dimensions" in str(e):
        ref_condition = re_encode_reference(width=width, height=height)
    else:
        raise

Prevention

Bind width/height to shared workflow fields used by both nodes.
Re-run the reference encoder whenever output resolution changes.
Don't hardcode resolution in one node and leave the other on template defaults.

When it happens

Trigger: Encoding a reference image at, say, 832x480 in the Reference Image node, then setting Wan Video Denoise width/height to 512x512 (or any other size).

Common situations: Changing output resolution in the denoise node after encoding, or using separate width/height fields in the two nodes that drifted apart; common when building workflows from templates with different defaults.

Related errors

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/10a262dd9c094fb1. Report an issue: GitHub.