Deoldify Vapoursynth filter

[Selur]

Please provide the pip command used to upgrade pytorch so that I can work on the same version.

First I used

python -m pip install torch==2.7.0.dev20250122+cu126

like WolframRhodium suggested over at https://github.com/AmusementClub/vs-dfttest2/issues/8

Now I want to setup a new Vapoursynth torch addon where I use (for vs-rife):

pip install -U packaging setuptools wheel
pip install --pre -U torch torchvision --index-url https://download.pytorch.org/whl/nightly/cu126
pip install --no-deps --pre -U torch_tensorrt --index-url https://download.pytorch.org/whl/nightly/cu126
pip install -U tensorrt-cu12 tensorrt-cu12_bindings tensorrt-cu12_libs --extra-index-url https://pypi.nvidia.com

which installs an even newer version. (see attachment for how I setup the torch add-on so far; still incomplete)

I cannot recompile the correlation sample at every torch build, please change it only if really necessary.

I guess there is no way to get rid of that dependency?

Since I from time to time need to update the torch-addon to support newer stuff, I need a way to get this working, ideally without you having to build it. If you could write down the instructions on how to build the correlation sample, I could do it myself whenever I need to setup a new torch-addon.

Alternatively, I would have to drop vsDeOldify since I can't install it.
I could set up a download for the last working Hybrid version with the torch-addon for that version on my filejump.com account, but I would not update vsDeOldify anymore.
(I could keep the current support for it in Hybrid, so users could still use newer versions, if the manage to add them and their dependencies to the Lib/site-packages as long as the vsDeOldify calls do not change.)

Cu Selur

Ps.: Please also check whether the Deoldify setup steps seem correct to you, thanks.