Using TLT with Kubernetes and Kubeflow

Hey everyone, our company just purchased a DGX system and I am playing around with it. We are trying to set up a Kubeflow pipeline utilizing the TLT. I am following along with a sample found here https://ngc.nvidia.com/catalog/model-scripts/nvidia:tlt_iva. However looking at the pipeline configuration I see that they are using the tlt-pull command to download a model. When I run the tlt-pull command I receive the following error ngc: error: unrecognized arguments: --apikey, --org, --team. Is the tlt-pull command outdated? Should I be using the ngc registry command to pull models? If so, how would I go about doing this in a Kubeflow pipeline manner. Any assistance would be great. I am brand new to all of this lol.

Hi Brandon,
Yes, the tlt-pull command is outdated.
Please refer to
https://devtalk.nvidia.com/default/topic/1063956/transfer-learning-toolkit/transfer-learning-toolkit-general-availability/
–>
https://developer.nvidia.com/transfer-learning-toolkit
–>
https://docs.nvidia.com/metropolis/
–>
https://docs.nvidia.com/metropolis/TLT/tlt-getting-started-guide/index.html

https://docs.nvidia.com/metropolis/TLT/tlt-getting-started-guide/index.html#downloading_models

Downloading a model

Use this command to download the model you have chosen from the NGC model registry:

ngc registry model download-version <ORG/model_name:version> -d <path_to_download_dir>