Kijai

/

LTXV2_comfy

like 426

What main model did you load? and did you connect the vae output of the main model loader somewhere?

Guessing from the error message you get saying : " the VAE is from a checkpoint loader node your checkpoint does not contain a valid VAE"

For video VAE I tested models like checkpoint, LTXV2 19b distilled fp8 . safetensor, it working well at setup, where I do not using audio VAE.
In another setup with audio encode, The message is addresed to Audio VAE and i have only one Audio VAE - LTX2_audio_vae_bf16.safetensor from this repository.
As is showed at screen grab, i use modue from KJ for audio vaeloaderKJ.
it is AUDIO VAE loader.

it is setup from Bejnis tutorials: LTX-2 Image2Video (Qwen3TTS) Ver. 20260127

LTXVAudioVAEEncode
ERROR: VAE is invalid: None
If the VAE is from a checkpoint loader node your checkpoint does not contain a valid VAE.

I take Audio vae from checkpoint of full model and it works fine. But separated Audio VAE don´t working.

One thing that is worth a try, make sure all is up to date. KJNodes specially. The loading of LTX-2 audio vae was added fairly recently.

Yes,I have up-date all nodes and frontend also backend of Comfy. KJnodes 1.2.9 ver. I simply fixed to use AudioVAE from Checkpoint loader. not Alone.
Thank you.

The odd thing is that from my limited understanding of error messages, it seems the error is from ComfyUI (and not KJNodes).
And from trying to use vae from the regular checkpoint loader

"File "D:\ComfyUI_windows_portable\ComfyUI\comfy_extras\nodes_lt_audio.py", line 54, in execute"
" VAE is from a checkpoint loader node your checkpoint does not contain a valid VAE."

But i'm no expert on that ;-)
And the KJnode vae loader works perfectly my end

Maybe one of the experts will see this and have a better idea

For video VAE I tested models like checkpoint, LTXV2 19b distilled fp8 . safetensor, it working well at setup, where I do not using audio VAE.

You either use KJ's models and the nodes ONLY or you use LTXV2- models and their nodes ONLY. Usually you can't mix it.

Just follow the workflow that comes with the screenshot from here (see first page: https://huggingface.co/Kijai/LTXV2_comfy/ )- not "from Bejnis tutorials: LTX-2 Image2Video (Qwen3TTS)"... another screenshot: either KJ's version with his models or LTXV's version with their model:

Hi,
I made a few tests with separate Audio VAE. I download testing project from RuneXX, where are located, I download all models, but still problem with Audio VAE.
Maybe it is something at my version of Comfy.
I downloaded fresh models files from repository, of course... :-)

got prompt
!!! Exception during processing !!! 'VAE' object has no attribute 'latent_frequency_bins'
Traceback (most recent call last):
File "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 530, 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 "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 334, 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 "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 308, in async_map_node_over_list
await process_inputs(input_dict, i)
File "D:\ComfyUI_windows_portable\ComfyUI\execution.py", line 296, in process_inputs
result = f(**inputs)
^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\comfy_api\internal_init.py", line 149, in wrapped_func
return method(locked_class, **inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\comfy_api\latest_io.py", line 1710, in EXECUTE_NORMALIZED
to_return = cls.execute(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI_windows_portable\ComfyUI\comfy_extras\nodes_lt_audio.py", line 153, in execute
audio_freq = audio_vae.latent_frequency_bins
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'VAE' object has no attribute 'latent_frequency_bins'

I have the exact same problem here, so we are two already. However I don't use the portable version of ComfyUI, I have the Desktop Version installed. So this can't be the reason, it's probably not ComfUI-related, I guess, but something different - perhaps some hidden conflict or incompatibility. I think I will go and ask Gemini or ChatGPT... Or else if somebody has found the source of this error meanwhile, please share your knowledge!

You probably need to switched to the nightly version of ComfyUI to get the most up to date changes.🤔