NVIDIA GPU-Optimized VMI is a virtual machine image tailored to accelerate workloads, encompassing machine learning, deep learning, data science, and high-performance computing (HPC). By deploying this VMI, you can swiftly provision a GPU-accelerated Compute Cloud VM with minimal effort. The image is based on Ubuntu and comes equipped with pre-installed NVIDIA GPU drivers, Docker, and NVIDIA container toolkit.
This VMI provides straightforward access to NVIDIA’s NGC Catalog, a repository housing GPU-optimized software. It enables you to effortlessly pull and run performance-tuned, rigorously tested, and NVIDIA-certified Docker containers. The NGC Catalog provides free access to containerized AI, Data Science, and HPC applications, pre-trained models, AI Software Development Kits (SDKs), and various resources that enable data scientists, developers, and researchers to focus on constructing and deploying solutions.
NVIDIA GPU-Optimized VMI has full certification from NVIDIA. It is optimized to deliver peak performance across an extensive spectrum of workloads on NVIDIA GPUs available in Nebius AI.
-
Click the button in this card to go to VM creation. The image will be automatically selected under Image/boot disk selection.
-
Under Computing resources, select a platform with GPUs.
-
Under Network settings, enable a public IP address for the VM (Public IP: Auto for a random address or List if you have a reserved static address).
-
Under Access, paste the public key from the pair into the SSH key field.
-
Create the VM.
-
To test the NVIDIA Container Toolkit, run a sample CUDA container:
sudo docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi
For details, see Running a Sample Workload in the NVIDIA Container Toolkit documentation. -
To test the NGC CLI, list container images available from NGC Catalog:
ngc registry image list
- Natural language processing (NLP) and translation.
- Working with images and videos: segmentation, synthesis, detecting objects, classifying and estimating facial features, emotions, and body poses.
- Automatic speech recognition (ASR, speech to text, STT), text to speech (TTS), and audio synthesis.
- Information retrieval-based question answering (IRQA).
- Healthcare applications: heart rate estimation, computational drug discovery, sequencing analysis etc.
Software | Version |
---|---|
Ubuntu | 22.04 LTS |
Docker | 24.0.2 |
Git | latest |
JupyterLab | latest |
Miniconda | latest |
NVIDIA Container Toolkit | 1.13.2-1 |
NVIDIA Data Center Driver | 535.161.08 |
NVIDIA NGC CLI | 3.29.0 |
Nebius AI CLI | latest |