Ngccli entrypoint fix


With regards to the issue of an ngccli update requiring a workaround in order to train TAO models detailed in (Tlt.components.docker_handler.docker_handler: Stopping container): Is there any update on a fix that will no longer require these workarounds? Will upgrading TAO or ngc to the newest versions fix it? The workaround is cumbersome and makes working with notebooks especially much more difficult, so any update on the status of a more permanent fix would be appreciated.


TAO: 3.21.11
NGC: 2.2.1

There is already a fix.

Please update the nvidia-tao.
$ pip3 install nvidia-tao==0.1.24

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.