Comfy-Org/ComfyUI_Nvidia_VFX_Nodes
1
0
Fork 0
mirror of https://github.com/Comfy-Org/ComfyUI_Nvidia_VFX_Nodes.git synced 2026-11-04 23:37:30 +00:00
Code Issues 31 Projects Packages Wiki Activity

[GH-ISSUE #4] nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12) #1

New issue
Open
opened 2026-10-26 22:27:40 +00:00 by bsakaguchi · 9 comments
bsakaguchi commented 2026-10-26 22:27:40 +00:00
Owner
Copy link

Originally created by @jtreminio on GitHub (Mar 10, 2026).
Original GitHub issue: https://github.com/Comfy-Org/ComfyUI_Nvidia_VFX_Nodes/issues/4

Using the PyTorch Release 26.01 container.

Trying to upscale an LTX2.3 video:

loaded completely; 5733.00 MB usable, 1384.94 MB loaded, full load: True
!!! Exception during processing !!! NvVFX_Load failed: The effect has not been properly initialized (code -12)
Traceback (most recent call last):
  File "/workspace/dlbackend/ComfyUI/execution.py", line 524, 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 "/workspace/dlbackend/ComfyUI/execution.py", line 333, 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 "/workspace/dlbackend/ComfyUI/execution.py", line 307, in _async_map_node_over_list
    await process_inputs(input_dict, i)
  File "/workspace/dlbackend/ComfyUI/execution.py", line 295, in process_inputs
    result = f(inputs)
             ^^^^^^^^^^^
  File "/workspace/dlbackend/ComfyUI/comfy_api/internal/init.py", line 149, in wrapped_func
    return method(locked_class, inputs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/workspace/dlbackend/ComfyUI/comfy_api/latest/_io.py", line 1764, in EXECUTE_NORMALIZED
    to_return = cls.execute(*args, **kwargs)
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/workspace/dlbackend/ComfyUI/custom_nodes/Nvidia_RTX_Nodes_ComfyUI/init.py", line 87, in execute
    sr.load()
  File "/usr/local/lib/python3.12/dist-packages/nvvfx/effects/base.py", line 119, in load
    self._effect.load()
nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12)

$ pip list | grep nvidia
nvidia-cuda-runtime-cu13               0.0.0a0
nvidia-cudnn-frontend                  1.17.0
nvidia-dali-cuda130                    1.53.0
nvidia-libnvcomp-cu13                  5.1.0.21
nvidia-ml-py                           13.590.44
nvidia-modelopt                        0.40.0
nvidia-nvimgcodec-cu13                 0.7.0.49
nvidia-nvjpeg                          13.0.2.28
nvidia-nvjpeg2k-cu13                   0.9.1.47
nvidia-nvtiff-cu13                     0.6.0.78
nvidia-resiliency-ext                  0.5.0
nvidia-vfx                             0.1.0.1

$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2025 NVIDIA Corporation
Built on Tue_Dec_16_07:23:41_PM_PST_2025
Cuda compilation tools, release 13.1, V13.1.115
Build cuda_13.1.r13.1/compiler.37061995_0

RTX Pro 6000

Originally created by @jtreminio on GitHub (Mar 10, 2026). Original GitHub issue: https://github.com/Comfy-Org/ComfyUI_Nvidia_VFX_Nodes/issues/4 Using the [PyTorch Release 26.01 container](https://docs.nvidia.com/deeplearning/frameworks/pytorch-release-notes/rel-26-01.html). Trying to upscale an LTX2.3 video: ``` loaded completely; 5733.00 MB usable, 1384.94 MB loaded, full load: True !!! Exception during processing !!! NvVFX_Load failed: The effect has not been properly initialized (code -12) Traceback (most recent call last): File "/workspace/dlbackend/ComfyUI/execution.py", line 524, 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 "/workspace/dlbackend/ComfyUI/execution.py", line 333, 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 "/workspace/dlbackend/ComfyUI/execution.py", line 307, in _async_map_node_over_list await process_inputs(input_dict, i) File "/workspace/dlbackend/ComfyUI/execution.py", line 295, in process_inputs result = f(inputs) ^^^^^^^^^^^ File "/workspace/dlbackend/ComfyUI/comfy_api/internal/init.py", line 149, in wrapped_func return method(locked_class, inputs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/dlbackend/ComfyUI/comfy_api/latest/_io.py", line 1764, in EXECUTE_NORMALIZED to_return = cls.execute(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/dlbackend/ComfyUI/custom_nodes/Nvidia_RTX_Nodes_ComfyUI/init.py", line 87, in execute sr.load() File "/usr/local/lib/python3.12/dist-packages/nvvfx/effects/base.py", line 119, in load self._effect.load() nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12) ``` ``` $ pip list | grep nvidia nvidia-cuda-runtime-cu13 0.0.0a0 nvidia-cudnn-frontend 1.17.0 nvidia-dali-cuda130 1.53.0 nvidia-libnvcomp-cu13 5.1.0.21 nvidia-ml-py 13.590.44 nvidia-modelopt 0.40.0 nvidia-nvimgcodec-cu13 0.7.0.49 nvidia-nvjpeg 13.0.2.28 nvidia-nvjpeg2k-cu13 0.9.1.47 nvidia-nvtiff-cu13 0.6.0.78 nvidia-resiliency-ext 0.5.0 nvidia-vfx 0.1.0.1 ``` ``` $ nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2025 NVIDIA Corporation Built on Tue_Dec_16_07:23:41_PM_PST_2025 Cuda compilation tools, release 13.1, V13.1.115 Build cuda_13.1.r13.1/compiler.37061995_0 ``` RTX Pro 6000
bsakaguchi commented 2026-10-26 22:27:43 +00:00
Author
Owner
Copy link

@yussufbiyik commented on GitHub (Mar 11, 2026):

!!! Exception during processing !!! NvVFX_Load failed: The effect has not been properly initialized (code -12)
Traceback (most recent call last):
  File "/root/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 "/root/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 "/root/ComfyUI/execution.py", line 308, in _async_map_node_over_list
    await process_inputs(input_dict, i)
  File "/root/ComfyUI/execution.py", line 296, in process_inputs
    result = f(**inputs)
             ^^^^^^^^^^^
  File "/root/ComfyUI/comfy_api/internal/__init__.py", line 149, in wrapped_func
    return method(locked_class, **inputs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/ComfyUI/comfy_api/latest/_io.py", line 1716, in EXECUTE_NORMALIZED
    to_return = cls.execute(*args, **kwargs)
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/ComfyUI/custom_nodes/comfyui_nvidia_rtx_nodes/__init__.py", line 87, in execute
    sr.load()
  File "/root/.local/lib/python3.12/site-packages/nvvfx/effects/base.py", line 119, in load
    self._effect.load()
nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12)

Prompt executed in 0.21 seconds

Happens on RTX 5060Ti too

<!-- gh-comment-id:4042173859 --> @yussufbiyik commented on GitHub (Mar 11, 2026): ``` !!! Exception during processing !!! NvVFX_Load failed: The effect has not been properly initialized (code -12) Traceback (most recent call last): File "/root/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 "/root/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 "/root/ComfyUI/execution.py", line 308, in _async_map_node_over_list await process_inputs(input_dict, i) File "/root/ComfyUI/execution.py", line 296, in process_inputs result = f(**inputs) ^^^^^^^^^^^ File "/root/ComfyUI/comfy_api/internal/__init__.py", line 149, in wrapped_func return method(locked_class, **inputs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/ComfyUI/comfy_api/latest/_io.py", line 1716, in EXECUTE_NORMALIZED to_return = cls.execute(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/ComfyUI/custom_nodes/comfyui_nvidia_rtx_nodes/__init__.py", line 87, in execute sr.load() File "/root/.local/lib/python3.12/site-packages/nvvfx/effects/base.py", line 119, in load self._effect.load() nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12) Prompt executed in 0.21 seconds ``` Happens on RTX 5060Ti too
bsakaguchi commented 2026-10-26 22:27:43 +00:00
Author
Owner
Copy link

@eclgo commented on GitHub (Mar 12, 2026):

Got the same error on RTX3080Ti Laptop GPU

<!-- gh-comment-id:4046391923 --> @eclgo commented on GitHub (Mar 12, 2026): Got the same error on RTX3080Ti Laptop GPU
bsakaguchi commented 2026-10-26 22:27:43 +00:00
Author
Owner
Copy link

@nregret commented on GitHub (Mar 14, 2026):

`!!! Exception during processing !!! NvVFX_Load failed: The effect has not been properly initialized (code -12)

Traceback (most recent call last):

File "/opt/ComfyUI/execution.py", line 524, 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 "/opt/ComfyUI/execution.py", line 333, 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 "/opt/ComfyUI/execution.py", line 307, in _async_map_node_over_list

await process_inputs(input_dict, i)

File "/opt/ComfyUI/execution.py", line 295, in process_inputs

result = f(**inputs)

File "/opt/ComfyUI/comfy_api/internal/init.py", line 149, in wrapped_func

return method(locked_class, **inputs)

File "/opt/ComfyUI/comfy_api/latest/_io.py", line 1764, in EXECUTE_NORMALIZED

to_return = cls.execute(*args, **kwargs)

File "/opt/ComfyUI/custom_nodes/comfyui_nvidia_rtx_nodes/init.py", line 87, in execute

sr.load()

~~~~~~~^^

File "/usr/local/lib64/python3.13/site-packages/nvvfx/effects/base.py", line 119, in load

self._effect.load()

~~~~~~~~~~~~~~~~~^^

nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12)`
4080 same questions

<!-- gh-comment-id:4059090357 --> @nregret commented on GitHub (Mar 14, 2026): `!!! Exception during processing !!! NvVFX_Load failed: The effect has not been properly initialized (code -12) Traceback (most recent call last): File "/opt/ComfyUI/execution.py", line 524, 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 "/opt/ComfyUI/execution.py", line 333, 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 "/opt/ComfyUI/execution.py", line 307, in _async_map_node_over_list await process_inputs(input_dict, i) File "/opt/ComfyUI/execution.py", line 295, in process_inputs result = f(**inputs) File "/opt/ComfyUI/comfy_api/internal/__init__.py", line 149, in wrapped_func return method(locked_class, **inputs) File "/opt/ComfyUI/comfy_api/latest/_io.py", line 1764, in EXECUTE_NORMALIZED to_return = cls.execute(*args, **kwargs) File "/opt/ComfyUI/custom_nodes/comfyui_nvidia_rtx_nodes/__init__.py", line 87, in execute sr.load() ~~~~~~~^^ File "/usr/local/lib64/python3.13/site-packages/nvvfx/effects/base.py", line 119, in load self._effect.load() ~~~~~~~~~~~~~~~~~^^ nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12)` 4080 same questions
bsakaguchi commented 2026-10-26 22:27:43 +00:00
Author
Owner
Copy link

@niaoyu123 commented on GitHub (Mar 14, 2026):

  1. 升级驱动版本到570.195.03+
  2. 容器环境变量配置NVIDIA_DRIVER_CAPABILITIES

你就能解决所有问题

<!-- gh-comment-id:4059907304 --> @niaoyu123 commented on GitHub (Mar 14, 2026): 1. 升级驱动版本到570.195.03+ 2. 容器环境变量配置NVIDIA_DRIVER_CAPABILITIES 你就能解决所有问题
bsakaguchi commented 2026-10-26 22:27:43 +00:00
Author
Owner
Copy link

@tillski791234 commented on GitHub (Mar 16, 2026):

It took me a half weekend.. This cryptic error is pure trolling from nvidia ;-) An intense strace helped to look for opened files and libs so I found it. This was an LXC node with passthrough but maybe the problem is also for docker.

Cause: missing libnvidia-ngx.so.1 inside the LXC container.

Why this was hard to find:

  • ldd looked clean
  • CUDA/Torch worked
  • failure happened only at NvVFX_Load / sr.load()

strace showed that NGX was loaded dynamically at runtime, and the container did not have:

/usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1

After bind-mounting/copying that library into the container and running ldconfig, nvvfx started working.

So if someone sees code -12 in a container, check for missing libnvidia-ngx.so.1, not just the usual CUDA/TensorRT libs.

I also discovered that nv-vfx is not working on a 5090. But on a 5060ti - so it´s not a general Blackwell problem. The output on the 5090 is always black. 4080 is also working. Driver: 590.44.01

<!-- gh-comment-id:4064555494 --> @tillski791234 commented on GitHub (Mar 16, 2026): It took me a half weekend.. This cryptic error is pure trolling from nvidia ;-) An intense strace helped to look for opened files and libs so I found it. This was an LXC node with passthrough but maybe the problem is also for docker. Cause: missing `libnvidia-ngx.so.1` inside the LXC container. Why this was hard to find: - `ldd` looked clean - CUDA/Torch worked - failure happened only at `NvVFX_Load` / `sr.load()` `strace` showed that NGX was loaded dynamically at runtime, and the container did not have: `/usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1` After bind-mounting/copying that library into the container and running `ldconfig`, `nvvfx` started working. So if someone sees `code -12` in a container, check for missing `libnvidia-ngx.so.1`, not just the usual CUDA/TensorRT libs. I also discovered that nv-vfx is not working on a 5090. But on a 5060ti - so it´s not a general Blackwell problem. The output on the 5090 is always black. 4080 is also working. Driver: 590.44.01
bsakaguchi commented 2026-10-26 22:27:44 +00:00
Author
Owner
Copy link

@jcp219 commented on GitHub (Mar 28, 2026):

For anybody else that gets this far and needs the extra connection between tillski791234's comment and the solution, see below.

You can find that file in your WSL folder through File Explorer:
\\wsl.localhost\Ubuntu\usr\lib\wsl\lib\libnvidia-ngx.so.1

Then either:

  • Copy it to your container (Replace DOCKER_CONTAINER_NAME with your Docker container name):
    docker cp \\wsl.localhost\Ubuntu\usr\lib\wsl\lib\libnvidia-ngx.so.1 DOCKER_CONTAINER_NAME:/usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1

  • Or add it to your docker folder and then update your dockerfile:
    COPY libnvidia-ngx.so.1 /usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1

PS: Thanks @tillski791234

<!-- gh-comment-id:4147351148 --> @jcp219 commented on GitHub (Mar 28, 2026): For anybody else that gets this far and needs the extra connection between tillski791234's comment and the solution, see below. You can find that file in your WSL folder through File Explorer: `\\wsl.localhost\Ubuntu\usr\lib\wsl\lib\libnvidia-ngx.so.1` Then either: - Copy it to your container (Replace DOCKER_CONTAINER_NAME with your Docker container name): `docker cp \\wsl.localhost\Ubuntu\usr\lib\wsl\lib\libnvidia-ngx.so.1 DOCKER_CONTAINER_NAME:/usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1` - Or add it to your docker folder and then update your dockerfile: `COPY libnvidia-ngx.so.1 /usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1` PS: Thanks @tillski791234
bsakaguchi commented 2026-10-26 22:27:44 +00:00
Author
Owner
Copy link

@Lolologist commented on GitHub (Apr 1, 2026):

Going from what jcp219 mentioned above, here's what got me unstuck (not in docker, just in WSL):
echo "/usr/lib/wsl/lib" | sudo tee /etc/ld.so.conf.d/wsl-nvidia.conf
sudo ldconfig

Alternatively, set the env var before launching ComfyUI:
export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH"

I had to open a new terminal to get that working and, VERY FRUSTRATINGLY, the first terminal I tried it in just silently died. New terminal after that and bam, working. No idea why that happened but at least there's a fix.

<!-- gh-comment-id:4172621495 --> @Lolologist commented on GitHub (Apr 1, 2026): Going from what jcp219 mentioned above, here's what got me unstuck (not in docker, just in WSL): `echo "/usr/lib/wsl/lib" | sudo tee /etc/ld.so.conf.d/wsl-nvidia.conf` `sudo ldconfig` Alternatively, set the env var before launching ComfyUI: `export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH"` I had to open a new terminal to get that working and, VERY FRUSTRATINGLY, the first terminal I tried it in just silently died. New terminal after that and bam, working. No idea why that happened but at least there's a fix.
bsakaguchi commented 2026-10-26 22:27:44 +00:00
Author
Owner
Copy link

@treilhes commented on GitHub (Sep 2, 2026):

Hello,
Just to provide a cleaner solution to this issue as copying the lib seems a bit dirty
By default the nvidia container toolkit does not provide all the libs to a container.
To configure the provided libs you need to configure your docker container with:

environment:
      NVIDIA_DRIVER_CAPABILITIES: all

You can also be more restrictive:

environment:
     NVIDIA_DRIVER_CAPABILITIES: compute,utility,graphics,video

Nvidia documentation:
https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/docker-specialized.html#driver-capabilities

<!-- gh-comment-id:5507008334 --> @treilhes commented on GitHub (Sep 2, 2026): Hello, Just to provide a cleaner solution to this issue as copying the lib seems a bit dirty By default the nvidia container toolkit does not provide all the libs to a container. To configure the provided libs you need to configure your docker container with: ``` environment: NVIDIA_DRIVER_CAPABILITIES: all ``` You can also be more restrictive: ``` environment: NVIDIA_DRIVER_CAPABILITIES: compute,utility,graphics,video ``` Nvidia documentation: https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/docker-specialized.html#driver-capabilities
bsakaguchi commented 2026-10-26 22:27:45 +00:00
Author
Owner
Copy link

@Donato-touch commented on GitHub (Sep 4, 2026):

FYI, RTX VSR worked correctly through the regular Comfy Cloud API, but I reproduced this error after packaging the same workflow as a Comfy Developer Platform Serverless Build.

  • GPU: RTX PRO 6000
  • ComfyUI: v0.34.2
  • CUDA 13.0 / Python 3.12 / PyTorch 2.12.1
  • nvidia-vfx: 0.1.0.1

The Serverless Build completed successfully, but RTX VSR failed at runtime with:

NvVFX_Load failed: The effect has not been properly initialized (code -12)

Regular CUDA inference and RealESRGAN worked on the same Serverless deployment.

<!-- gh-comment-id:5540845257 --> @Donato-touch commented on GitHub (Sep 4, 2026): FYI, RTX VSR worked correctly through the regular Comfy Cloud API, but I reproduced this error after packaging the same workflow as a Comfy Developer Platform Serverless Build. - GPU: RTX PRO 6000 - ComfyUI: v0.34.2 - CUDA 13.0 / Python 3.12 / PyTorch 2.12.1 - nvidia-vfx: 0.1.0.1 The Serverless Build completed successfully, but RTX VSR failed at runtime with: `NvVFX_Load failed: The effect has not been properly initialized (code -12)` Regular CUDA inference and RealESRGAN worked on the same Serverless deployment.
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
alexis/rename-true-hdr-to-video-hdr
Revert-node-name
alexis/add-long-video-true-hdr-workflow
alexis/add-node-thumbnails-readme
alexis/add-video-upscale-thumbnail
alexis/update-workflow-thumbnails
alexis/update-workflows
Update-README
alexis/update-tru-hdr-workflow
alexis/update-vfx-nodes
alexis/update-pyproject
add-publish-workflow-and-pyproject
No results found.
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
Comfy-Org/ComfyUI_Nvidia_VFX_Nodes#1
No description provided.