As part of RHODS-2978, `jupyter-server-proxy` was added to Tensorflow and Pytorch imaged. However to display Tensorboard, users need to manually add TENSORBOARD_PROXY_URL env variable to access Tensorboard as follows:
import os
os.environ["TENSORBOARD_PROXY_URL"]= os.environ["NB_PREFIX"]+"/proxy/6006/"
Add TENSORBOARD_PROXY_URL if not present as part of spawning a Tensorflow or Pytorch notebook.