### Install TAO Launcher CLI

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/quick_start_guide/running_via_launcher.html

Use the quick start script to install the TAO launcher and NGC CLI.

```bash
bash setup/quickstart_launcher.sh --install
```

--------------------------------

### Install TAO CLI

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_deploy/nsight_dl_designer_integration.html

Setup a Python virtual environment and install the TAO client library.

```bash
# Create and activate virtual environment
python3 -m venv tao_env
source tao_env/bin/activate

# Install TAO
pip install nvidia-tao-client
```

--------------------------------

### Install and Start Minikube

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_deployment.html

Installs Minikube for local Kubernetes development and starts a cluster with Docker driver, Docker runtime, and GPU support. It also exposes port 32080.

```bash
#install minikube
curl -LO https://github.com/kubernetes/minikube/releases/latest/download/minikube-linux-amd64
sudo install minikube-linux-amd64 /usr/local/bin/minikube && rm minikube-linux-amd64

#start minikube
minikube start --driver=docker --container-runtime=docker --gpus=all --ports=32080:32080
```

--------------------------------

### Environment Setup for Workflow Example

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_inference_microservice.html

Set up environment variables for API base URL, organization name, and NGC API key. This script also handles authentication to obtain an access token.

```bash
# Environment setup
export BASE_URL="http://<your-api-host>:<port>/api/v2"
export NGC_ORG="your-org-name"
export NGC_API_KEY="your-ngc-api-key"

# Authenticate
TOKEN=$(curl -s -X POST $BASE_URL/login \
    -H "Content-Type: application/json" \
    -d '{
        "ngc_key": "'"$NGC_API_KEY"'",
        "ngc_org_name": "'"$NGC_ORG"'"
    }' | jq -r '.token')

echo "Authenticated with token"

# Set experiment and training job IDs (from previous training)
EXPERIMENT_ID="your-experiment-id"
TRAIN_JOB_ID="your-completed-training-job-id"


```

--------------------------------

### Object Detection Workflow Example

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_remote_client.html

An example of initiating an object detection workflow using the v2 API. This includes fetching base experiments, getting training schemas, and setting up the initial training job configuration. Remember to edit the spec file as needed.

```bash
# Complete object detection workflow with v2 API
export WORKSPACE_ID="workspace_789"
export DATASET_URI="<scheme>://my-bucket/coco-dataset"  # aws://, azure://, lustre://, file://, seaweedfs://

# 1. Get base experiment and training schema
tao rtdetr list-base-experiments --filter-param network_arch=rtdetr --output json | jq .
export SELECTED_PTM_ID="ptm_id_from_list"

# Get training schema and save to file
tao rtdetr get-job-schema --action train --base-experiment-id "$SELECTED_PTM_ID" --output @train_spec.yaml
# Edit train_spec.yaml as needed. For example, update "epochs": 100 and "learning_rate": 0.001

```

--------------------------------

### Navigate to Setup Directory

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_docker_compose.html

Change the working directory to the Docker Compose setup folder.

```bash
cd tao_tutorials/setup/tao-docker-compose
```

--------------------------------

### Launch NSight DL Designer

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_deploy/nsight_dl_designer_integration.html

Commands to navigate to the installation directory and start the NSight DL Designer application.

```bash
cd <NDLD_INSTALL_DIR>/host/linux-desktop-dl-x64
./nsight-dl
```

--------------------------------

### Example Analyze Command (TAO Launcher)

Source: https://docs.nvidia.com/tao/tao-toolkit/text/data_services/data_analytics.html

An example of how to run the data analyze command with a default experiment specification file.

```bash
tao dataset analytics analyze -e $DEFAULT_SPEC
```

--------------------------------

### Execute Prune Command Examples

Source: https://docs.nvidia.com/tao/tao-toolkit/text/cv_finetuning/tensorflow_2/image_classification_tf2.html

Specific examples for triggering the pruning process via FTMS or the TAO Launcher.

```bash
PRUNE_JOB_ID=$(tao classification_tf2 create-job \
  --kind experiment \
  --name "classification_tf2_prune" \
  --action prune \
  --workspace-id $WORKSPACE_ID \
  --parent-job-id $TRAIN_JOB_ID \
  --specs @prune_spec.yaml \
  --base-experiment-id "$BASE_EXPERIMENT_ID" \
  --encryption-key "nvidia_tlt" \
  --output json | jq -r '.id')
```

```bash
tao model classification_tf2 prune -e /workspace/spec.yaml
```

--------------------------------

### TAO Launcher Prune Command Example

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/tensorflow_2/image_classification_tf2.html

Example of executing the prune command using the TAO Launcher.

```bash
tao model classification_tf2 prune -e /workspace/spec.yaml
```

--------------------------------

### Install TAO Client Versions

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/ds_tao/deepstream_tao_integration.html

Commands to install specific versions of the TAO client in a virtual environment.

```bash
pip3 install nvidia-pyindex
pip3 install nvidia-tao-client==0.1.19    # for 3.0-21.08
pip3 install nvidia-tao-client==0.1.20    # for 3.0-21.11
```

--------------------------------

### Install Docker Engine

Source: https://docs.nvidia.com/tao/tao-toolkit/text/tao_toolkit_api/api_setup.html

Configures the Docker repository and installs the Docker engine, CLI, and plugins on Ubuntu.

```bash
# Add Docker's official GPG key:
sudo apt-get update
sudo apt-get install ca-certificates curl
sudo install -m 0755 -d /etc/apt/keyrings
sudo curl -fsSL https://download.docker.com/linux/ubuntu/gpg -o /etc/apt/keyrings/docker.asc
sudo chmod a+r /etc/apt/keyrings/docker.asc

# Add the repository to Apt sources:
echo \
"deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.asc] https://download.docker.com/linux/ubuntu \
$(. /etc/os-release && echo "${UBUNTU_CODENAME:-$VERSION_CODENAME}") stable" | \
sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
sudo apt-get update

#install docker
sudo apt-get install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin

#add user to docker group
sudo usermod -aG docker $USER

newgrp docker
```

--------------------------------

### Start Triton Server

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/pytorch/panoptic_3d_reconstruction/nvpanoptix3d.html

Execute this command to start the Triton Inference Server. Ensure you are in the correct directory.

```bash
bash scripts/nvpanoptix3d_e2e_inference/start_server.sh
```

--------------------------------

### Set Up Environment Variables

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_airgapped_deployment.html

Install the required package and configure environment variables for model downloading.

```bash
# Install tao-core package
pip install -U nvidia-tao-core

# Set environment variables
export AIRGAPPED_MODE=True
export PTM_API_KEY="your_ngc_api_key"
```

--------------------------------

### Clone TAO Tutorials Repository

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/quick_start_guide/index.html

Use this command to download the TAO tutorials repository, which contains setup scripts and tutorial notebooks for TAO.

```bash
git clone https://github.com/NVIDIA/tao_tutorials.git
```

--------------------------------

### Start Inference Microservice via cURL

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_inference_microservice.html

Example command to start an inference microservice using cURL and capture the job ID.

```bash
INFERENCE_MS_JOB_ID=$(curl -s -X POST \
    $BASE_URL/orgs/$NGC_ORG/inference_microservices:start \
    -H "Authorization: Bearer $TOKEN" \
    -H "Content-Type: application/json" \
    -d '{
        "parent_job_id": "'"$TRAIN_JOB_ID"'"
    }' | jq -r '.job_id')

echo "Inference Microservice Job ID: $INFERENCE_MS_JOB_ID"
```

--------------------------------

### Start Jupyter Notebook

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/quick_start_guide/running_via_launcher.html

Launch the Jupyter notebook server to begin working with TAO tutorials.

```bash
jupyter notebook --ip 0.0.0.0 --port 8888 --allow-root
```

--------------------------------

### Install Git LFS for DeepStream Apps

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/ds_tao/deformable_detr_ds.html

Required setup for cloning the DeepStream TLT apps repository.

```bash
curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | sudo bash
sudo apt-get install git-lfs
git lfs install
```

--------------------------------

### Initialize DeepStream Development Environment

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/ds_tao/nvocdr_ds.html

Commands to start the appropriate DeepStream container and install OpenCV dependencies.

```bash
docker run --gpus=all -v <work_path>:<work_path> --rm -it --privileged --net=host nvcr.io/nvidia/deepstream:6.2-triton bash

# install opencv
apt update && apt install -y libopencv-dev
```

```bash
docker run --gpus=all -v <work_path>:<work_path> --rm -it --privileged --net=host nvcr.io/nvidia/deepstream-l4t:6.2-triton bash

# install opencv
apt update && apt install -y libopencv-dev
```

```bash
# install opencv
apt update && apt install -y libopencv-dev
```

--------------------------------

### Example Sparse4D Training Command with TAO Launcher

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/pytorch/sparse4d/sparse4d.html

An example command for initiating Sparse4D training using TAO Launcher, specifying the experiment specification file and results directory.

```bash
tao model sparse4d train -e $DEFAULT_SPEC results_dir=$RESULTS_DIR

```

--------------------------------

### Sample TAO EfficientDet Training Command

Source: https://docs.nvidia.com/tao/tao-toolkit/text/cv_finetuning/tensorflow_2/efficientdet_tf2.html

Example of how to initiate the training process for an EfficientDet model using the TAO launcher, specifying the experiment specification file and the number of GPUs to use.

```bash
tao model efficientdet_tf2 train -e /path/to/spec.yaml num_gpus=2

```

--------------------------------

### Example TAO Help Output

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/quick_start_guide/running_via_launcher.html

Sample output displayed when running the help command for the TAO launcher.

```text
usage: tao [-h] {list,stop,info,dataset,deploy,model} ...

Launcher for TAO

optional arguments:
-h, --help            show this help message and exit

task_groups:
      {list,stop,info,dataset,deploy,model}
```

--------------------------------

### Example Training Status Output

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/pytorch/metric_learning_recognition/metric_learning_recognition.html

Sample JSON output from the status.json file, indicating the start and successful completion of the training process.

```json
{"date": "6/20/2023", "time": "23:11:2", "status": "STARTED", "verbosity": "INFO", "message": "Starting Training Loop."}
...

{"date": "6/20/2023", "time": "23:11:22", "status": "SUCCESS", "verbosity": "INFO", "message": "Train finished successfully."}
```

--------------------------------

### Sample Evaluation Status JSON

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_deploy/metric_learning_recognition.html

An example of the status.json file generated during evaluation, showing the start and success status of the ml_recog evaluation process.

```json
{"date": "3/30/2023", "time": "6:7:14", "status": "STARTED", "verbosity": "INFO", "message": "Starting ml_recog evaluation."}
{"date": "3/30/2023", "time": "6:7:24", "status": "SUCCESS", "verbosity": "INFO", "message": "Evaluation finished successfully."}
```

--------------------------------

### Download and Run Jupyter Notebooks

Source: https://docs.nvidia.com/tao-toolkit-api/hosted-ea/jupyter_notebook_examples.html

Use these commands to download the TAO API notebooks, install necessary packages, and start the Jupyter Notebook server.

```bash
wget https://api-ea4.tao.ngc.nvidia.com/tao_api_notebooks.zip
sudo apt-get update
sudo apt-get install unzip
unzip tao_api_notebooks.zip
pip3 install jupyter notebook
jupyter notebook --ip 0.0.0.0
```

--------------------------------

### Object Detection Workflow

Source: https://docs.nvidia.com/tao/tao-toolkit/text/tao_toolkit_api/api_remote_client.html

An example of initiating an object detection workflow using the v2 API. This snippet shows how to get base experiments and training schemas, similar to the classification workflow.

```bash
# Complete object detection workflow with v2 API
export WORKSPACE_ID="workspace_789"
export DATASET_URI="<scheme>://my-bucket/coco-dataset"  # aws://, azure://, lustre://, file://, seaweedfs://

# 1. Get base experiment and training schema
tao rtdetr list-base-experiments --filter-param network_arch=rtdetr --output json | jq .
export SELECTED_PTM_ID="ptm_id_from_list"

# Get training schema and save to file
tao rtdetr get-job-schema --action train --base-experiment-id "$SELECTED_PTM_ID" --output @train_spec.yaml
# Edit train_spec.yaml as needed. For example, update "epochs": 100 and "learning_rate": 0.001
```

--------------------------------

### Sample ~/.tao_mounts.json configuration

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_launcher.html

A complete example of the configuration file defining mounts, environment variables, and Docker options.

```json
{
    "Mounts": [
        {
            "source": "/path/to/your/data",
            "destination": "/workspace/tao-experiments/data"
        },
        {
            "source": "/path/to/your/local/results",
            "destination": "/workspace/tao-experiments/results"
        },
        {
            "source": "/path/to/your/local/specs",
            "destination": "/workspace/tao-experiments/specs"
        }
    ],
    "Envs": [
        {
            "variable": "CUDA_DEVICE_ORDER",
            "value": "PCI_BUS_ID"
        }
    ],
    "DockerOptions": {
        "shm_size": "16G",
        "ulimits": {
            "memlock": -1,
            "stack": 67108864
        },
        "user": "1000:1000",
        "ports": {
            "8888": "8888"
        }
    }
}
```

--------------------------------

### Sample Usage for TAO Launcher Evaluation

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_deploy/deformable_detr.html

An example command demonstrating how to run evaluation with the TAO launcher, specifying default variables for the spec file, results directory, and engine file.

```bash
tao deploy deformable_detr evaluate -e $DEFAULT_SPEC
        results_dir=$RESULTS_DIR \
        evaluate.trt_engine=$ENGINE_FILE

```

--------------------------------

### Get Base Experiment ID and Training Specs

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/pytorch/depth_estimation/monocular_depth_estimation.html

Retrieves the base experiment ID and default training specifications for a monocular depth estimation model using the FTMS client. Ensure jq is installed for JSON parsing.

```bash
# Get the training spec
BASE_EXPERIMENT_ID=$(tao monocular_depth_estimation list-base-experiments | jq -r '.[0].id')
TRAIN_SPECS=$(tao monocular_depth_estimation get-job-schema --action train --base-experiment-id $BASE_EXPERIMENT_ID | jq -r '.default')
```

--------------------------------

### TAO Launcher Training with Overrides

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/embedding/clip/training.html

Example of running CLIP model training with TAO Launcher, demonstrating how to specify the experiment file and override parameters such as the number of GPUs and the results directory.

```bash
tao model clip train -e /path/to/experiment_spec.yaml \
  train.num_gpus=4 \
  train.num_epochs=50 \
  results_dir=/results/clip_run1
```

--------------------------------

### Sample SSD Model Training Command

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/tensorflow_1/object_detection/ssd.html

An example of how to execute the SSD model training command, specifying the number of GPUs, the experiment specification file, the results directory, and the encryption key.

```bash
tao model ssd train --gpus 2 -e /path/to/spec.txt -r /path/to/result -k $KEY
```

--------------------------------

### Install NVIDIA GPU Driver

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_setup.html

Installs NVIDIA GPU driver version 570 on Ubuntu. A reboot is required after installation.

```bash
#install nv driver
sudo apt-get update && sudo apt-get install nvidia-driver-570
sudo reboot

```

--------------------------------

### Install Helm 3

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_deployment.html

Installs Helm version 3 on your system. This script downloads and executes the official Helm installation script.

```bash
#install helm
curl -fsSL -o get_helm.sh https://raw.githubusercontent.com/helm/helm/main/scripts/get-helm-3
chmod 700 get_helm.sh
./get_helm.sh
```

--------------------------------

### Sample EfficientDet Training Command

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/tensorflow_2/efficientdet_tf2.html

A basic example of executing the training command with an experiment specification and GPU configuration.

```bash
tao model efficientdet_tf2 train -e /path/to/spec.yaml num_gpus=2
```

--------------------------------

### Start the Triton Server

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/pytorch/pose_classification/pose_classification.html

Initializes the Triton inference server to enable engine generation.

```bash
bash scripts/start_server.sh
```

--------------------------------

### Install nvidia-tao-client Package

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/faqs.html

Use these commands to install the latest TAO client package. Uninstall the previous `nvidia-tlt` if it was previously installed in your virtual environment.

```bash
pip3 uninstall nvidia-tlt       # This is only required if you had previously installed nvidia-tlt in your
                                # Virtual environment.
pip3 install nvidia-tao-client

```

--------------------------------

### Sample Training Command

Source: https://docs.nvidia.com/tao/tao-toolkit/text/cv_finetuning/pytorch/object_detection/grounding_dino.html

Example usage of the training command specifying an experiment configuration file.

```bash
tao grounding_dino model train -e /path/to/spec.yaml
```

--------------------------------

### Install IPython Kernel for Conda Environment

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/quick_start_guide/running_via_launcher.html

Install the ipykernel package and then install the kernel for the 'launcher' Conda environment. This ensures notebooks use the correct Python kernel.

```bash
python -m pip install ipykernel
python -m ipykernel install --user --name launcher --display-name "launcher"

```

--------------------------------

### Evaluation Configuration Example

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/pytorch/image_classification_pyt.html

Example configuration file for model evaluation.

```yaml
evaluate:
  checkpoint: /path/to/model.pth
```

--------------------------------

### Verify GPU Operator Installation

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_deployment.html

Checks if the GPU operator pods are running and lists installed Helm charts. This command is useful for verifying the installation on various Kubernetes platforms.

```bash
kubectl get pods -n gpu-operator

helm list
```

--------------------------------

### Example: Run Annotation Slice Command

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/data_services/annotations.html

An example demonstrating how to use the `slice` command in Data Services with a specified experiment file.

```bash
tao dataset annotations slice -e /path/to/spec.yaml
```

--------------------------------

### Install NVIDIA GPU Operator

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_toolkit_api/api_deployment.html

Installs the NVIDIA GPU Operator using Helm. This command adds the NVIDIA Helm repository, updates it, and then installs the operator in the 'gpu-operator' namespace.

```bash
#install gpu operator
helm repo add nvidia https://helm.ngc.nvidia.com/nvidia \
   && helm repo update

helm install --wait --generate-name \
   -n gpu-operator --create-namespace \
   nvidia/gpu-operator \
   --version=v25.3.0 \
   --set driver.enabled=false \
   --set toolkit.enabled=false
```

--------------------------------

### Install TensorBoard

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tensorboard_visualization.html

Use pip to install the TensorBoard package in your Python environment.

```bash
python -m pip install tensorboard
```

--------------------------------

### Example: Running SSD Inference

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/tensorflow_1/object_detection/ssd.html

An example of using the 'inference' command with SSD models. Ensure all required paths and the model key are correctly provided.

```bash
tao model ssd inference -i /path/to/input/images_dir -o /path/to/output/dir -m /path/to/trained_tlt_ssd_model -k <model_key> -e /path/to/ssd_spec.txt
```

--------------------------------

### Sample Evaluation Command (TAO Launcher)

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/tensorflow_2/efficientdet_tf2.html

Example of how to use the `evaluate` command with the TAO launcher, specifying the experiment specification file.

```bash
tao model efficientdet_tf2 evaluate -e /path/to/spec.yaml
```

--------------------------------

### Example Pruning Command

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/tensorflow_1/image_classification.html

Example usage of the pruning command with specific parameters.

```bash
tao model classification_tf1 prune -m /workspace/output/weights/resnet_003.tlt
                         -o /workspace/output/weights/resnet_003_pruned.tlt
                         -eq union
                         -pth 0.7 -k $KEY
```

--------------------------------

### Install AWS CLI

Source: https://docs.nvidia.com/tao/tao-toolkit/text/tao_toolkit_api/api_setup.html

Downloads and installs the AWS Command Line Interface for Linux.

```bash
#install aws cli
curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
unzip awscliv2.zip
sudo ./aws/install
```

--------------------------------

### Create New Experiment

Source: https://docs.nvidia.com/tao-toolkit-api/hosted-ea/tao_rest_api.html

Initializes a new experiment using specified network architecture, datasets, and cloud configuration.

```bash
EXPERIMENT_ID=$(curl -s -X POST \
 $BASE_URL/orgs/ea-tlt/experiments \
 -H "Content-Type: application/json" \
 -H "Authorization: Bearer $TOKEN" \
 -d '{
   "network_arch": "dino",
   "encryption_key": "tlt_encode",
   "checkpoint_choose_method": "best_model",
   "train_datasets": ["'"$TRAIN_DATASET_ID"'"],
   "eval_dataset": "'"$EVAL_DATASET_ID"'",
   "inference_dataset": "'"$EVAL_DATASET_ID"'",
   "calibration_dataset": "'"$TRAIN_DATASET_ID"'",
   "docker_env_vars": {},
   "base_experiment": ["'"$BASE_EXPERIMENT_ID"'"],
   "cloud_details": {
       "cloud_type": "aws",
       "cloud_specific_details": {
           "cloud_region": "us-west-1",
           "cloud_bucket_name": "bucket_name",
           "access_key": "access_key",
           "secret_key": "secret_key"
       }
   }
}' | jq -r '.id')
echo $EXPERIMENT_ID
```

--------------------------------

### Install TAO Deploy via Pip

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao_deploy/tao_deploy_overview.html

Install the nvidia-tao-deploy package in a Python environment.

```bash
pip install nvidia-tao-deploy
```

--------------------------------

### Example of Using the Prune Command

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/cv_finetuning/tensorflow_1/semantic_segmentation/unet.html

An example demonstrating how to use the `prune` task with specific parameters for equalization and pruning threshold. The key is provided as a shell variable.

```bash
tao model unet prune -e </path/to/spec/file>
               -m </path/to/weights to be pruned>
               -o </path/to/pruned weights>
               -eq union
               -pth 0.7
               -k $KEY

```

--------------------------------

### Install OpenSSL Dependency

Source: https://docs.nvidia.com/tao/tao-toolkit/6.26.03/text/tao-converter/tao_converter_deformable_detr.html

Install the required OpenSSL development package on the target machine.

```bash
sudo apt-get install libssl-dev
```