# NIM

Serve optimized model containers with NIM in a Flyte task.

[NVIDIA NIM](https://www.nvidia.com/en-in/ai/), part of NVIDIA AI Enterprise, provides a streamlined path
for developing AI-powered enterprise applications and deploying AI models in production.
It includes an out-of-the-box optimization suite, enabling AI model deployment across any cloud,
data center, or workstation. Since NIM can be self-hosted, there is greater control over cost, data privacy,
and more visibility into behind-the-scenes operations.

With NIM, you can invoke the model's endpoint as if it is hosted locally, minimizing network overhead.

## Installation

To use the NIM plugin, run the following command:

```shell
$ pip install flytekitplugins-inference
```

## Example usage

For a usage example, see [NIM example usage](https://www.union.ai/docs/v1/flyte/integrations/flytekit-plugins/nim-plugin/serve-nim-container/page.md).

> [!NOTE]
> NIM can only be run in a Flyte cluster as it must be deployed as a sidecar service in a Kubernetes pod.

## Subpages

- [Serve Generative AI Models with NIM](https://www.union.ai/docs/v1/flyte/integrations/flytekit-plugins/nim-plugin/serve-nim-container/page.md)

---
**Source**: https://github.com/unionai/unionai-docs/blob/main/content/integrations/flytekit-plugins/nim-plugin/_index.md
**HTML**: https://www.union.ai/docs/v1/flyte/integrations/flytekit-plugins/nim-plugin/