# max-nvidia-base

max-nvidia-base is a self-hosted ai app. MAX is Modular's self-hosted platform for deploying and running large language model (LLM) inference endpoints on your own hardware. It is open source and self-hosted, installable on Docker Hub.

Canonical page: https://ourlocalhost.com/app/6448/max-nvidia-base

## Description

MAX is Modular's self-hosted platform for deploying and running large language model (LLM) inference endpoints on your own hardware. This lightweight NVIDIA base container keeps dependencies minimal for faster downloads, though it requires NVIDIA GPU (graphics processing unit) drivers for hardware acceleration.

## Platform availability

Available on 0 of 10 platforms.

## Health score

23/100 (Low), based on 2 of 4 factors.
- Easy to install: 10 (0 install platforms, Docker image)
- Light to run: 40 (4 GB minimum RAM, runs on ARM)

## Links

- Docker Hub: https://hub.docker.com/r/modular/max-nvidia-base

## Common questions

### Is max-nvidia-base a full environment?

It is the MAX base container and provides a lightweight environment with minimal dependencies.

### Does the image include CUDA?

Yes, CUDA is listed as one of the essential components included in the container.

### What PyTorch variant is included?

The container includes PyTorch (CPU).

### Is cuDNN included in the base image?

No, cuDNN is omitted because the image avoids heavier frameworks.

---
Source: Our Local Host (https://ourlocalhost.com/app/6448/max-nvidia-base), the self-hosted and open-source app directory.
