invoke-ai/InvokeAI · error · ValueError

No Qwen3 Encoder source provided. Standalone safetensors/GGU

Error message

No Qwen3 Encoder source provided. Standalone safetensors/GGUF models require a separate text encoder. Options:
  1. Set 'Qwen3 Encoder' to a standalone Qwen3 text encoder model (Klein 4B needs Qwen3 4B, Klein 9B needs Qwen3 8B)
  2. Set 'Qwen3 Source' to a Diffusers Flux2 Klein model to extract the encoder from

What it means

The FLUX.2 Klein model loader raised ValueError because the standalone safetensors/GGUF main model was given no Qwen3 text encoder source — neither a standalone Qwen3 encoder model nor a Diffusers Flux2 Klein pipeline to extract it from. Standalone checkpoints do not include the text encoder, so the loader cannot build a runnable model.

Source

Thrown at invokeai/app/invocations/flux2_klein_model_loader.py:180

            )

        # Determine Qwen3 Encoder source
        if self.qwen3_encoder_model is not None:
            # Use standalone Qwen3 Encoder - validate it matches the FLUX.2 Klein variant
            self._validate_qwen3_encoder_variant(context, main_config)
            qwen3_tokenizer = self.qwen3_encoder_model.model_copy(update={"submodel_type": SubModelType.Tokenizer})
            qwen3_encoder = self.qwen3_encoder_model.model_copy(update={"submodel_type": SubModelType.TextEncoder})
        elif main_is_diffusers:
            # Extract from main model (recommended for FLUX.2 Klein)
            qwen3_tokenizer = self.model.model_copy(update={"submodel_type": SubModelType.Tokenizer})
            qwen3_encoder = self.model.model_copy(update={"submodel_type": SubModelType.TextEncoder})
        elif self.qwen3_source_model is not None:
            # Extract from separate Diffusers model
            self._validate_encoder_source(context, self.qwen3_source_model, "Qwen3 Source", main_config)
            qwen3_tokenizer = self.qwen3_source_model.model_copy(update={"submodel_type": SubModelType.Tokenizer})
            qwen3_encoder = self.qwen3_source_model.model_copy(update={"submodel_type": SubModelType.TextEncoder})
        else:
            raise ValueError(
                "No Qwen3 Encoder source provided. Standalone safetensors/GGUF models require a separate text encoder. "
                "Options:\n"
                "  1. Set 'Qwen3 Encoder' to a standalone Qwen3 text encoder model "
                "(Klein 4B needs Qwen3 4B, Klein 9B needs Qwen3 8B)\n"
                "  2. Set 'Qwen3 Source' to a Diffusers Flux2 Klein model to extract the encoder from"
            )

        return Flux2KleinModelLoaderOutput(
            transformer=TransformerField(transformer=transformer, loras=[]),
            qwen3_encoder=Qwen3EncoderField(tokenizer=qwen3_tokenizer, text_encoder=qwen3_encoder),
            vae=VAEField(vae=vae),
            max_seq_len=self.max_seq_len,
        )

    def _validate_diffusers_format(
        self, context: InvocationContext, model: ModelIdentifierField, model_name: str
    ) -> AnyModelConfig:
        """Validate that a model exposes the diffusers-style submodel layout and return its config.

View on GitHub (pinned to 0b6a024f2f)

Solutions

Set 'Qwen3 Encoder' to a standalone Qwen3 text encoder model (Qwen3 4B for Klein 4B, Qwen3 8B for Klein 9B)
Set 'Qwen3 Source' to a Diffusers FLUX.2 Klein pipeline so tokenizer+encoder are extracted from it
Install the appropriate Qwen3 encoder via the model manager
Catch ValueError and prompt the user to supply one of the two encoder sources

Example fix

// before
loader = Flux2KleinModelLoader(model=safetensors_checkpoint, vae_model=flux_vae)
// after
loader = Flux2KleinModelLoader(
    model=safetensors_checkpoint,
    vae_model=flux_vae,
    qwen3_encoder_model=qwen3_4b_encoder,   # Klein 4B -> Qwen3 4B
    # or: qwen3_source_model=diffusers_klein_pipeline
)
output = loader.invoke(context)

Defensive patterns

Strategy: validation

Validate before calling

cfg = context.models.get_config(loader.model)
if cfg.format != ModelFormat.Diffusers and loader.qwen3_encoder_model is None and loader.qwen3_source_model is None:
    raise ValueError("standalone checkpoint needs Qwen3 encoder or Qwen3 Source")

Type guard

def has_encoder_source(loader) -> bool:
    return loader.qwen3_encoder_model is not None or loader.qwen3_source_model is not None

Try / catch

try:
    output = loader.invoke(context)
except ValueError as e:
    if 'No Qwen3 Encoder source provided' in str(e):
        loader.qwen3_encoder_model = pick_qwen3_encoder_for(loader.model)
        output = loader.invoke(context)
    else:
        raise

Prevention

Pair the encoder size to the Klein size (4B->Qwen3 4B, 9B->Qwen3 8B)
Install both standalone FLUX VAE and Qwen3 encoders alongside single-file checkpoints
Prefer Diffusers pipelines that bundle tokenizer and encoder

When it happens

Trigger: invoke() reaches encoder resolution with self.model in safetensors/GGUF format while self.qwen3_encoder_model is None and self.qwen3_source_model is None (final else branch).

Common situations: Single-file Klein checkpoint downloaded without the required Qwen3 encoder; user unaware that Klein 4B needs Qwen3 4B and Klein 9B needs Qwen3 8B; workflow saved before the encoder input was wired.

Related errors

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/96f9e4e2b2b9765f. Report an issue: GitHub.