Getting Access#

Requesting a TensorRT Cloud Enabled NGC Org#

TensorRT-Cloud is in Early Access (EA). To secure tenancy, file a DevZone request: https://developer.nvidia.com/tensorrt-cloud-program.

A TensorRT-Cloud admin will review your request and account creation instructions will be sent to your email upon approval.

Prerequisites

Ensure you’re using Python ≥ 3.10 or < 3.13.
To perform TensorRT refit, the tensorrt or tensorrt-lean Python package must be installed with the version matching the TensorRT engine. The tensorrt-lean package is required for refitting version-compatible engines (those built with the --vc trtexec flag), and tensorrt is required for refitting non-version-compatible engines.
```
$ python3 -m pip install tensorrt==10.0.1 tensorrt-cu12==10.0.1
$ python3 -m pip install tensorrt-lean==10.0.1 tensorrt_lean-cu12==10.0.1
```
For more information, refer to the NVIDIA TensorRT Installation Guide.
If you would like to refit a TensorRT-LLM engine, install it locally. For more information, refer to the NVIDIA TensorRT-LLM Quick Start Guide.

Note

refit is not supported on MacOS as TensorRT does not support MacOS.

Steps

The TensorRT-Cloud CLI tool is distributed via PyPi: https://pypi.org/project/trt-cloud/

Use pip to install.
```
$ python3 -m pip install trt-cloud
```
Verify the TensorRT-Cloud CLI tool was installed successfully.
```
$ trt-cloud --version
```
Get an overview of the CLI usage.
```
trt-cloud --help
```

Prerequisites

Steps

Log into https://ngc.nvidia.com.
Click Setup under the user profile.
Click Generate API Key.
Generate a personal key with the NVIDIA TensorRT Cloud Service and the Private Registry services enabled.
Log into the TensorRT-Cloud CLI.
```
$ trt-cloud login --nvapi-key <GENERATED_TOKEN>
```
Important
- We recommend that you rotate your API key frequently for security best practices.
- Your API key is stored as plain text on your machine under ~/.trt-cloud/config.