micromamba-docker: error: NVTX functions not installed. Are you sure you have a CUDA build?

Hi all, me again.

I am using image FROM mambaorg/micromamba:focal-cuda-11.8.0 but PyTorch has an issue:

NVTX functions not installed. Are you sure you have a CUDA build

Tried adding via apt-get

    && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends apt-utils  \
        build-essential \
        cmake \
        cuda-command-line-tools-$(cut -f1,2 -d- <<< ${CUDA//./-}) \
        git \
        wget \
        curl \
        nano \
        nvidia-cuda-toolkit \

And added to env via yaml.

name: SE3nv
channels:
#  - defaults
  - conda-forge
  - pytorch
  - dglteam
  - nvidia
dependencies:
  - python=3.9
  - pytorch=1.9
  - torchaudio
  - torchvision
  - cudatoolkit=11.1
  - dgl-cuda11.1
  - nvtx
  - icecream
  - pip
  - pip:
    - hydra-core
    - pyrsistent

Any advise to fix?

Thanks,

Lars