TensorFlow throws missing libdevice errors

[yankcrime]

Using TensorFlow results in the following error:

2024-01-08 21:30:47.213628: W external/local_xla/xla/service/gpu/llvm_gpu_backend/gpu_backend_lib.cc:504] Can't find libdevice directory ${CUDA_DIR}/nvvm/libdevice. This may result in compilation or runtime failures, if the program we try to run uses routines from libdevice.
Searched for CUDA in the following directories:
  ./cuda_sdk_lib
  /usr/local/cuda-12.2
  /usr/local/cuda
  /opt/conda/lib/python3.11/site-packages/tensorflow/python/platform/../../../nvidia/cuda_nvcc
  /opt/conda/lib/python3.11/site-packages/tensorflow/python/platform/../../../../nvidia/cuda_nvcc

I've tested this with v1.5_cuda-12.0_ubuntu-22.04 as well as building locally from master and it's easily reproducible by following the TensorBoard Getting Started example.