NameError: name 'pipe' is not defined

[CS6700]

Hi, just setup VisionDepth3D today with conda, it launches fine.
However when I try to use the depth estimation tool I get this:

Using a slow image processor as use_fast is unset and a slow processor was saved with this model. use_fast=True will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with use_fast=False.
⬇️ Downloaded model from Hugging Face: xingyang1/Distill-Any-Depth-Large-hf
Device set to use cuda:0
🔥 Hugging Face pipeline warmed up with dummy frame
Exception in thread Thread-2 (process_video2):
Traceback (most recent call last):
File "C:\Users\shimi\miniconda3\envs\VD3D\Lib\threading.py", line 1075, in _bootstrap_inner
self.run()
File "C:\Users\shimi\miniconda3\envs\VD3D\Lib\threading.py", line 1012, in run
self._target(*self._args, **self._kwargs)
File "C:\Users\shimi\Desktop\VisionDepth3D-main\core\render_depth.py", line 782, in process_video2
if hasattr(pipe, "image_processor") and hasattr(pipe.image_processor, "export_depth_to_16bit_png"):
^^^^
NameError: name 'pipe' is not defined

On top of that, it does state that CUDA is available and running on CUDA at startup, then why is it using ['CPUExecutionProvider'] and cuda:0?

Python 3.12
CUDA 12.8
Pythorch 2.7.0

4060Ti OC 16GB
Ryzen 5 X7600
32GB of ram

Thanks in advance.

[VisionDepth]

sorry for the late response, i thought i updated the repo with the patch for the pipe call, must still be in the 3.2.4 branch under branches, will upload the new update tmrw, onnx providers running cpu is just letting you know that the rife onnx upscale models couldn't find tensorRT in your system so it fell back to CPU, totally fine still runs on cpu just in the Frametools tab, when loading the safetensor from hugging face its letting you know that its running on device 0 meaning that your gpu is running, if you had multiple GPUs it would tell you which GPU it was running off of, since you only have one it brings back 0, everything was normal except the pipe not being loaded, that's when i implemented marigold diffusion and forgot to tie back in the hugging face pipeline and onnx model logic,

[CS6700]

Thanks a lot for clarifying things.
So all I need to do is update VisionDepth3D tomorrow?

[VisionDepth]

I have now updated the repo, to the most recent updates, inference logic for all depth models should work properly now, Depthcrafter is Still a WIP

https://github.com/VisionDepth/VisionDepth3D#%EF%B8%8F-guide-sheet-updating-visiondepth3d