Remote Client - NVIDIA Docs

The remote client allows you to create model experiments using the command line instead of relying on API calls.

Installation

Copy
Copied!

            
            pip3 install -y nvidia-tao-client

Storage Topology

Below is a detectnet_v2 example of shared storage content.

Copy
Copied!

            
            shared
└── users
    ├── 1a40ddb6-d389-4c39-9b40-eaa948ee508d
    │   ├── metadata.json
    │   ├── datasets
    │   │   ├── 78cff931-e867-4e9c-b8a7-30fa243e433c
    │   │   │   ├── metadata.json
    │   │   │   ├── specs
    │   │   │   │   └── convert.json
    │   │   │   ├── images
    │   │   │   │   ├── 000000.jpg
    │   │   │   │   └── …
    │   │   │   ├── labels
    │   │   │   │   ├── 000000.txt
    │   │   │   │   └── …
    │   │   │   ├── 444fed7a-ac3f-4499-8a57-0234a43c752d
    │   │   │   │   ├── status.json
    │   │   │   │   └── …
    │   │   │   └── tfrecords
    │   │   │       └── …
    │   │   └── …
    │   └── models
    │       ├── cdf387d2-067e-4935-be9d-7facbc851f60
    │       │   ├── metadata.json
    │       │   ├── specs
    │       │   │   ├── train.json
    │       │   │   └── …
    │       │   ├── 444fed7a-ac3f-4499-8a57-0234a43c754e
    │       │   │   ├── status.json
    │       │   │   └── …
    │       │   └── experiment_dir_unpruned
    │       │       └── …
    │       └── …
    └── …

CLI Specs

The command line depends on the following environment variables:

BASE_URL
USER
TOKEN

The command line is composed of:

Copy
Copied!

            
            $ tao-client <network> <action> <args>

For example:

Copy
Copied!

            
            $ tao-client detectnet-v2 run-action --id 042559ec-ab3e-438d-9c94-2cab38f76efc --action dataset_convert --job_type dataset

Note

You can always use –help argument to retrieve the command usage information.

To list supported networks:

Copy
Copied!

            
            $ tao-client –help

To list supported detectnet-v2 actions:

Copy
Copied!

            
            $ tao-client detectnet-v2 –help

Examples

The tutorial notebooks provide examples that demonstrate the various workflows.

Download the resource using the NGC CLI.

Copy
Copied!

            
            ngc registry resource download-version "nvidia/tao/tao-getting-started:5.3.0"

Find tutorial notebooks (.ipynb files).

Copy
Copied!

            
            cd tao-getting-started_v5.3.0/notebooks/tao_api_starter_kit/client

Instead of using NGC, the tutorial notebooks can be downloaded from the machine where TAO API is deployed. You can obtain them using:

Copy
Copied!

            
            wget https://<ip_of_hosted_machine>:<nginx_service_forwarded_port>/tao_api_notebooks.zip
unzip tao_api_notebooks.zip
cd cli

Serve these Notebook examples using Jupyter.

Copy
Copied!

            
            pip3 install jupyter notebook
jupyter notebook --ip 0.0.0.0