DownloadAndLoadHyVideoTextEncoder: argument of type 'NoneType' is not iterable

[Azureth86]

The workflow I'm trying to do is pretty much the example workflow:
https://github.com/kijai/ComfyUI-HunyuanVideoWrapper/blob/main/example_workflows/hyvideo_leapfusion_img2vid_example_01.json

I can see the folders for the LLM, so I don't know what could be missing.

got prompt
Loading text encoder model (clipL) from: C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\models\clip\clip-vit-large-patch14
Text encoder to dtype: torch.float16
Loading tokenizer (clipL) from: C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\models\clip\clip-vit-large-patch14
Loading text encoder model (llm) from: C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\models\LLM\llava-llama-3-8b-text-encoder-tokenizer
!!! Exception during processing !!! argument of type 'NoneType' is not iterable
Traceback (most recent call last):
File "C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\execution.py", line 349, in execute
output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\execution.py", line 224, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\timo\ComfyUI_windows_portable\ComfyUI\execution.py", line 196, in map_node_over_list
process_inputs(input_dict, i)
File "C:\Users\timo\ComfyUI_windows_portable\ComfyUI\execution.py", line 185, in process_inputs
results.append(getattr(obj, func)(**inputs))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper\nodes.py", line 740, in loadmodel
text_encoder = TextEncoder(
^^^^^^^^^^^^
File "C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper\hyvideo\text_encoder_init_.py", line 180, in init
self.model, self.model_path = load_text_encoder(
^^^^^^^^^^^^^^^^^^
File "C:\Users\timo_\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper\hyvideo\text_encoder_init_.py", line 41, in load_text_encoder
text_encoder = AutoModel.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\timo_\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\models\auto\auto_factory.py", line 571, in from_pretrained
return model_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\timo_\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\modeling_utils.py", line 309, in wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\timo\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\modeling_utils.py", line 4507, in from_pretrained
model = cls(config, *model_args, **model_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\timo_\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\models\llama\modeling_llama.py", line 379, in init
self.post_init()
File "C:\Users\timo_\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\modeling_utils.py", line 1968, in post_init
if v not in ALL_PARALLEL_STYLES:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: argument of type 'NoneType' is not iterable

[kijai]

I recommend using the bridge node and ComfyUI native text encoding as it has far less issues than the original transformers based text encoder loading:

[Azureth86]

Thanks.

Though I had to downgrade from the 1.0.8 to the 'nightly' version, since even though the 1.0.8 had the bridge node in the extension node list, it had no preview and I didn't find it in my comfyUI node list.

Now I'm at OOM, even though I can run the basic hunyuan workflow. I have 'only' 16G VRAM and 32 RAM. Will continue fiddling.

Edit: Hmm, the auto_cpu_offload seems to work for me, but honestly I'm just dabbling around half blind.

[Azureth86]

For anyone with same problem, the original issue seems to be me having a comfyUI portable that was set to use +cu118 CUDA drivers. Once I did clean install with PyTorch 2.7+cu128 I'm at least moving forward.