HunyuanVideo gets stuck when loading video encoder

[gioxyer]

Starting server

To see the GUI go to: http://127.0.0.1:8188
FETCH DATA from: C:\Users\giorg\Documents\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager\extension-node-map.json [DONE]
got prompt
The config attributes {'mid_block_causal_attn': True} were passed to AutoencoderKLCausal3D, but are not expected and will be ignored. Please verify your config.json configuration file.
encoded latents shape torch.Size([1, 16, 11, 96, 96])
Loading text encoder model (clipL) from: C:\Users\giorg\Documents\ComfyUI_windows_portable\ComfyUI\models\clip\clip-vit-large-patch14
Text encoder to dtype: torch.float16
Loading tokenizer (clipL) from: C:\Users\giorg\Documents\ComfyUI_windows_portable\ComfyUI\models\clip\clip-vit-large-patch14
C:\Users\giorg\Documents\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\tokenization_utils_base.py:1601: FutureWarning: clean_up_tokenization_spaces was not set. It will be set to True by default. This behavior will be depracted in transformers v4.45, and will be then set to False by default. For more details check this issue: huggingface/transformers#31884
warnings.warn(
Loading text encoder model (llm) from: C:\Users\giorg\Documents\ComfyUI_windows_portable\ComfyUI\models\LLM\llava-llama-3-8b-text-encoder-tokenizer
Loading checkpoint shards: 100%|████████████████████████████████████████████████████████| 4/4 [12:41<00:00, 190.36s/it]

[kijai]

How much memory do you have? It looks like not enough, if you update the nodes there should be quantization option for the text encoder, it requires bitsandbytes installed but reduces the memory use of the text encoder by a lot.

[gioxyer]

I have 8gb of VRAM, in which part of text encoder is there that options? Thanks

[kijai]

You should have this option

For 8gb you would also need to use the block_swap -feature, check the example workflow for more info.

[gioxyer]

I have these on your nodes

[kijai]

Then the nodes are not up to date.

[gioxyer]

I reinstall from manager but gives this issue now

[raultresd]

I have the same problem and I don't think it´s a memory lack problem in my case.

[raultresd]

I solved the problem updating the dependencies, which only updated torch to 2.5.1