fix(build): remove non-redistributable CUDA files from CUDA and XPU container images

[dabhi365]

📝 Description

Two issues where non-redistributable NVIDIA CUDA files were not being correctly removed from published container images, violating the NVIDIA EULA.

XPU image (#3597):
The XPU build pipeline branches off of anomalib-studio-base and has no cleanup stage. The triton dependencies mentioned in the issue are not even used by Triton as they have an Intel backend. This fix removes it after uv sync.

CUDA Image (#3598):
The existing cleanup had an extra flag -path "*cupti*" for the find command. The non-compliant files are installed via pip wheels and land under .../nvidia/cu13/lib/ so the filter was skipping them and they survived into the final image.

• Fixes Remove non-redistributable CUDA files from XPU image #3597
• Fixes Remove non-redistributable CUDA files from CUDA image #3598

✨ Changes

Select what type of change your PR is:

• 📦 Build system changes
• 🔒 Security update

✅ Checklist

Before you submit your pull request, please make sure you have completed the following steps:

• 📚 I have made the necessary updates to the documentation (if applicable).
• 🧪 I have written tests that support my changes and prove that my fix is effective or my feature works (if applicable).
• 🏷️ My PR title follows conventional commit format.

Note:
The CUDA fix was checked by running the container on my device, but I was not able to test and confirm the XPU fix due to lack of hardware.