Traceback ConstructorError(None, None)

[revolvedai]

Running on Linux, Set cuda device to: 0

Total VRAM 24241 MB, total RAM 25033 MB

pytorch version: 2.1.0+cu118

xformers version: 0.0.22.post4+cu118

Set vram state to: NORMAL_VRAM

Device: cuda:0 NVIDIA RTX A5000 : native

VAE dtype: torch.bfloat16

Using xformers cross attention

---

[0.5922, 0.5647, 0.5333]]]])" --driving_audio "/opt/rd/apps/ComfyUI/input/audio.wav" --output /opt/rd/apps/ComfyUI/output/hallo_1718727686477756178.mp4 --pose_weight 1.0 --face_weight 1.0 --lip_weight 1.0 --face_expand_ratio 1.2

Traceback (most recent call last):

File "/opt/rd/apps/ComfyUI/custom_nodes/ComfyUI-Hallo/Hallo/scripts/inference.py", line 372, in

inference_process(command_line_args)

File "/opt/rd/apps/ComfyUI/custom_nodes/ComfyUI-Hallo/Hallo/scripts/inference.py", line 128, in inference_process

config = OmegaConf.load(args.config)

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/omegaconf/omegaconf.py", line 188, in load

obj = yaml.load(f, Loader=get_yaml_loader())

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/init.py", line 81, in load

return loader.get_single_data()

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/constructor.py", line 51, in get_single_data

return self.construct_document(node)

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/constructor.py", line 60, in construct_document

for dummy in generator:

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/constructor.py", line 413, in construct_yaml_map

value = self.construct_mapping(node)

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/omegaconf/_utils.py", line 151, in construct_mapping

return super().construct_mapping(node, deep=deep)

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/constructor.py", line 218, in construct_mapping

return super().construct_mapping(node, deep=deep)

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/constructor.py", line 143, in construct_mapping

value = self.construct_object(value_node, deep=deep)

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/constructor.py", line 100, in construct_object

data = constructor(self, node)

File "/opt/rd/miniconda3/envs/CMFY/lib/python3.10/site-packages/yaml/constructor.py", line 427, in construct_undefined

raise ConstructorError(None, None,

yaml.constructor.ConstructorError: could not determine a constructor for the tag 'tag:yaml.org,2002:python/object/apply:torch._utils._rebuild_tensor_v2'

in "/opt/rd/apps/ComfyUI/custom_nodes/ComfyUI-Hallo/tmp.yaml", line 37, column 15

Prompt executed in 12.23 seconds

[AIFSH]

img or audio filename pure en not include Chinese，try agin

[revolvedai]

No chinese, I used a.png for image and audio.wav for audio. Same issue. Freezes at 50% progress. Then same Constructor Errors.

[gaoxuxu110]

I have the same issue: File "H:\Deepfacelab\stablediffusion\ComfyUI-aki-v1.1_test.ext\Lib\site-packages\yaml\constructor.py", line 427, in construct_undefined raise ConstructorError(None, None, yaml.constructor.ConstructorError: could not determine a constructor for the tag 'tag:yaml.org,2002:python/object/apply:torch._utils._rebuild_tensor_v2' in "H:\Deepfacelab\stablediffusion\ComfyUI-aki-v1.1_test\custom_nodes\ComfyUI-Hallo\tmp.yaml", line 37, column 15

504 [HalloNode]: 9.33s

by the way, no chinese in image and audio too

[mpoplawskiatpp]

your errors are the same as in Issue #2 Which make five of us having the same critical error :( Unfortunately I can't figure out what causes this.

[AIFSH]

LoadImagePath node not LoadImage

[mpoplawskiatpp]

OMG.... you are right. Wow. Is it possible that everybody from 5 people having this problem have the same temporary blindness like me :) :) :) a gigantic RED mark around node name would help :)

[revolvedai]

I have been using LoadImagePath and am still having the issue. I recreated the workflow exactly while testing.

[AIFSH]

I have been using LoadImagePath and am still having the issue. I recreated the workflow exactly while testing.

rename your audio filename，no "(1)" such character in it

[zdaar]

@revolvedai Maybe like me you were also using the wrong LoadAudioPath node. The one from ComfyUI AudioScheduler doesnt work with this workflow

[revolvedai]

Okay I figured it out. I was using the correct nodes. It was that the resolution of my image was too big. When I switched to 512x it worked. Could we perhaps add better error handling for resolution related errors so that the end user can realize what needs to be done? Thank you for this awesome node!

[revolvedai]

Do you need to use the Load Image Path and Load Audio Path node? There are so many conflicts in ComfyUI with those nodes so if you send someone a workflow to load, it shows all the conflicting nodes as necessary. That means the user downloads all the conflicting nodes too! Maybe some work for resolving the nodes or otherwise (ugh) making HalloLoadImagePath and HalloLoadAudioPath or something.