krea2-turbo-iQ3_xxs_im.gguf
Hey, how to use this model in ComfyUI?
Regular Unet Loader (GGUF) has problems loading it:
[ERROR] !!! Exception during processing !!! This model is not currently supported - (Unknown model architecture!)
[ERROR] Traceback (most recent call last):
File "C:\Development\ComfyUI\custom_nodes\ComfyUI-GGUF\loader.py", line 103, in gguf_sd_loader
arch_str = detect_arch(set(val[0] for val in tensors)).arch
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Development\ComfyUI\custom_nodes\ComfyUI-GGUF\tools\convert.py", line 169, in detect_arch
assert model_arch is not None, "Unknown model architecture!"
^^^^^^^^^^^^^^^^^^^^^^
AssertionError: Unknown model architecture!
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "c:\Development\ComfyUI\execution.py", line 542, in execute
output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "c:\Development\ComfyUI\execution.py", line 341, in get_output_data
return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "c:\Development\ComfyUI\execution.py", line 315, in _async_map_node_over_list
await process_inputs(input_dict, i)
File "c:\Development\ComfyUI\execution.py", line 303, in process_inputs
result = f(**inputs)
^^^^^^^^^^^
File "C:\Development\ComfyUI\custom_nodes\ComfyUI-GGUF\nodes.py", line 169, in load_unet
sd, extra = gguf_sd_loader(unet_path)
^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Development\ComfyUI\custom_nodes\ComfyUI-GGUF\loader.py", line 105, in gguf_sd_loader
raise ValueError(f"This model is not currently supported - ({e})")
ValueError: This model is not currently supported - (Unknown model architecture!)
I should probably add a model card to explain this, but all my image model GGUFs are intended to be used with sd.cpp, which uses weights fingerprinting rather than metadata to identify the model's architecture, unlike the ComfyUI-GGUF node, which expects metadata to know the model architecture. It has a fallback to fingerprinting for some architectures, but apparently not for Krea-2.
I probably won't bother dong it myself to fix it for everyone, but you could use HuggingFace's buit-in GGUF editor to fix the metadata, or the devs of ComfyUI-GGUF might implement fingerprinting for trhis arch.