invoke-ai/InvokeAI · error · TypeError

Expected ModelPatchRaw for LoRA '{lora.lora.key}', got {type

Error message

Expected ModelPatchRaw for LoRA '{lora.lora.key}', got {type(lora_info.model).__name__}. The LoRA model may be corrupted or incompatible.

What it means

When applying LoRAs to the Qwen3 text encoder, _lora_iterator loads each LoRA model and requires the loaded object to be a ModelPatchRaw (InvokeAI's raw patch representation). Any other loaded type means the file is not a usable Anima LoRA patch, so it raises a TypeError suggesting the model is corrupted or incompatible.

Source

Thrown at invokeai/app/invocations/anima_text_encoder.py:215

        context.util.signal_progress("Tokenizing with T5-XXL")
        t5_tokenizer = load_bundled_t5_tokenizer()
        t5_tokens = t5_tokenizer(
            prompt,
            padding=False,
            truncation=True,
            max_length=T5_MAX_SEQ_LEN,
            return_tensors="pt",
        )
        t5xxl_ids = t5_tokens.input_ids[0]  # Shape: (seq_len,)

        return qwen3_embeds, t5xxl_ids, None

    def _lora_iterator(self, context: InvocationContext) -> Iterator[PatchSpec]:
        """Iterate over LoRA models to apply to the Qwen3 text encoder."""
        for lora in self.qwen3_encoder.loras:
            lora_info = context.models.load(lora.lora)
            if not isinstance(lora_info.model, ModelPatchRaw):
                raise TypeError(
                    f"Expected ModelPatchRaw for LoRA '{lora.lora.key}', got {type(lora_info.model).__name__}. "
                    "The LoRA model may be corrupted or incompatible."
                )
            yield (lora_info.model, lora.weight, lora_info.model_in_ram())

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Re-import the LoRA with the correct model type (LoRA patch) in the model manager.
2. Replace the LoRA file with a fresh download and re-scan models.
3. Remove the LoRA from the qwen3_encoder loras list if it is not a Qwen3/Anima-compatible patch.

from invokeai.backend.model_manager.load import ModelPatchRaw
info = context.models.load(lora.lora)
if not isinstance(info.model, ModelPatchRaw):
    print(f"{lora.lora.key} is {type(info.model).__name__}, not a LoRA patch")

from invokeai.backend.model_manager.load import ModelPatchRaw
def is_lora_patch(info) -> bool:
    return isinstance(info.model, ModelPatchRaw)

try:
    result = invocation.invoke(context)
except TypeError as e:
    if "Expected ModelPatchRaw for LoRA" in str(e):
        reimport_lora_with_correct_model_type(e)
    else:
        raise

When it happens

Trigger: During invoke → _encode_prompt → _lora_iterator, when context.models.load(lora.lora).model is not an instance of ModelPatchRaw — e.g. the model record was imported with the wrong model type or the file is not a valid LoRA patch.

Common situations: Importing a checkpoint or full-model file as a LoRA; a LoRA trained for another base leaking into the qwen3_encoder list with mismatched metadata; corrupted download that parses as a different object type.

Related errors

• Expected ModelPatchRaw for LoRA '{lora.lora.key}', got {type
• LoRA "{lora_key}" already applied to transformer.
• LoRA "{lora_key}" already applied to Qwen3 encoder.
• Unknown lora: {lora.lora.key}!
• LoRA '{lora.lora.key}' is for {lora.lora.base.value if lora.

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/c07685fa69b95052. Report an issue: GitHub.