invoke-ai/InvokeAI · error · TypeError

Expected AutoencoderKLWan for Wan VAE, got {type(vae_info.mo

Error message

Expected AutoencoderKLWan for Wan VAE, got {type(vae_info.model).__name__}.

What it means

invoke() loads the VAE via context.models.load and asserts it is a diffusers AutoencoderKLWan instance, because the decode path depends on Wan-specific APIs (config.z_dim, Wan decode semantics). Any other VAE class raises this TypeError with the actual class name.

Source

Thrown at invokeai/app/invocations/wan_latents_to_video.py:107

    @torch.no_grad()
    def invoke(self, context: InvocationContext) -> VideoOutput:
        latents = context.tensors.load(self.latents.latents_name)
        _validate_video_latent_batch(latents)
        if latents.ndim == 4:
            # Promote 4D (single-frame) to 5D so this node can also serve as a
            # one-frame "video" encode if someone wires it that way.
            latents = latents.unsqueeze(2)
        if latents.ndim != 5:
            raise ValueError(
                f"Wan latents-to-video expects a 5D latent tensor [B, C, T, H, W]; got {tuple(latents.shape)}."
            )
        if any(size == 0 for size in latents.shape[2:]):
            raise ValueError("Wan latents-to-video requires non-empty temporal and spatial dimensions.")

        vae_info = context.models.load(self.vae.vae)
        if not isinstance(vae_info.model, AutoencoderKLWan):
            raise TypeError(f"Expected AutoencoderKLWan for Wan VAE, got {type(vae_info.model).__name__}.")

        if latents.shape[1] != vae_info.model.config.z_dim:
            raise ValueError(
                f"Latent channel mismatch: these latents have {latents.shape[1]} channels but the "
                f"selected VAE expects {vae_info.model.config.z_dim}. A14B models need the 16-channel Wan 2.1 VAE; "
                "TI2V-5B needs the 48-channel Wan 2.2 VAE."
            )

        _, _, t_lat, h_lat, w_lat = latents.shape
        spatial_scale = getattr(vae_info.model.config, "scale_factor_spatial", None) or 8
        temporal_scale = getattr(vae_info.model.config, "scale_factor_temporal", None) or 4
        t_pixel = (t_lat - 1) * temporal_scale + 1
        h_pixel, w_pixel = h_lat * spatial_scale, w_lat * spatial_scale
        optimize_memory = context.config.get().wan_memory_optimization

        estimated_working_memory = estimate_vae_working_memory_wan(
            operation="decode",
            vae=vae_info.model,

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Point the node's vae field at a Wan VAE model (AutoencoderKLWan).
2. Check the loaded model type in the Model Manager and correct the record if it resolves to the wrong class.
3. Create a fresh VAE node referencing the Wan 2.1/2.2 VAE instead of reusing an image-model VAE.

Example fix

// before
vae = sdxl_vae  # AutoencoderKL
video = wan_latents_to_video(latents=latents, vae=vae)  # TypeError
// after
vae = wan_vae  # AutoencoderKLWan (Wan 2.1 or 2.2)
video = wan_latents_to_video(latents=latents, vae=vae)

vae_info = context.models.load(vae.vae)
if not isinstance(vae_info.model, AutoencoderKLWan):
    raise TypeError(f"Need AutoencoderKLWan, got {type(vae_info.model).__name__}")

def is_wan_vae(model) -> bool:
    return isinstance(model, AutoencoderKLWan)

try:
    video = node.invoke(context)
except TypeError as e:
    if "Expected AutoencoderKLWan" in str(e):
        vae = load_wan_vae_for(transformer_variant)
        video = replace(node, vae=vae).invoke(context)
    else:
        raise

When it happens

Trigger: Wiring a non-Wan VAE (e.g. SDXL AutoencoderKL, AutoencoderKLWan wrapped differently, Flux VAE) into the vae field of wan_latents_to_video and calling invoke().

Common situations: Reusing a VAE node from an SD/Flux workflow template; a model manager record resolving to the wrong model class; copy-pasting a VAE model ID between workflows.

Related errors

• {type(config).__name__} is a single-file config; it does not
• Expected Main_GGUF_Wan_Config, got {type(config).__name__}.
• Expected Main_Checkpoint_Wan_Config, got {type(config).__nam
• Expected LlavaOnevisionForConditionalGeneration, got {type(m
• Expected AutoencoderKLWan or FluxAutoEncoder for Anima VAE,

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/56c1c6d30fb4890e. Report an issue: GitHub.