ModuleNotFoundError: No module named 'ltx_pipelines.multigpu'

[kylezhanzhe-fs]

(LTX-2) (base) root@autodl-container-119b47bef5-0081acce:~/autodl-tmp/20260516/LTX-2# python -m ltx_pipelines.ti2vid_two_stages
--checkpoint-path /root/autodl-tmp/ComfyUI/models/checkpoints/ltx-2.3-22b-dev.safetensors
--distilled-lora /root/autodl-tmp/ComfyUI/models/loras/ltx-2.3-22b-distilled-lora-384-1.1.safetensors 0.8
--spatial-upsampler-path /root/autodl-tmp/ComfyUI/models/latent_upscale_models/ltx-2.3-spatial-upscaler-x2-1.1.safetensors
--gemma-root /root/autodl-tmp/gemma-3-12b-it
--prompt "A beautiful sunset over the ocean"
--output-path output.mp4
Traceback (most recent call last):
File "", line 189, in _run_module_as_main
File "", line 112, in _get_module_details
File "/root/autodl-tmp/20260516/LTX-2/packages/ltx-pipelines/src/ltx_pipelines/init.py", line 15, in
from ltx_pipelines.a2vid_two_stage import A2VidPipelineTwoStage
File "/root/autodl-tmp/20260516/LTX-2/packages/ltx-pipelines/src/ltx_pipelines/a2vid_two_stage.py", line 15, in
from ltx_pipelines.utils.args import default_2_stage_arg_parser
File "/root/autodl-tmp/20260516/LTX-2/packages/ltx-pipelines/src/ltx_pipelines/utils/init.py", line 1, in
from ltx_pipelines.utils.blocks import (
File "/root/autodl-tmp/20260516/LTX-2/packages/ltx-pipelines/src/ltx_pipelines/utils/blocks.py", line 69, in
from ltx_pipelines.multigpu.delegating_builder import DelegatingBuilder
ModuleNotFoundError: No module named 'ltx_pipelines.multigpu'

[jpschw]

I'm getting the same problem after recent updates. This happens even in single-GPU mode, and is not specific to the default setup—it's a breaking change in the latest LTX-2 release it seems. I'm attempting to create a patch for myself. If it doesn't work then I will attempt a downgrade.

[kylezhanzhe-fs]

It seems they forgot to update certain files. Releasing code that doesn't even run properly is really irresponsible.

[jpschw]

How is it irresponsible? Just ask Claude to fix it until they do. I got it working after 5min. They've made an amazing model. Treat them with respect.

[imxtx]

How is it irresponsible? Just ask Claude to fix it until they do. I got it working after 5min. They've made an amazing model. Treat them with respect.

Have you fixed it? I'm struggling with the same issue. 🥹

[CrescentVelvet]

I encountered the same issue, so I created an empty file to replace it.

[kylezhanzhe-fs]

If this problem is caused by incomplete update files, modifying the code ourselves to fix it may not ensure that there are no other hidden risks. It would be better for the manufacturer to update the entire code.