What Vantage Compute Works With
The software Vantage Compute runs and connects to: NVIDIA Run:ai, NeMo, NIM and Dynamo, KServe, Kubeflow, Ray, JupyterLab, Spark, and VAST, WEKA and DDN storage.
GPU orchestration
NVIDIA Run:ai
Deploy NVIDIA Run:ai into a Kubernetes cluster that Vantage Compute manages, and open the Run:ai interface from the Vantage Compute interface.
Training
NVIDIA NeMo
Train and fine-tune models with NVIDIA NeMo, on the GPUs you already have.
Kubeflow
Run Kubeflow on your own GPU clusters, under the same identity as every other job.
Ray
Run distributed Ray workloads on your own GPU clusters.
Serving
NVIDIA NIM
Serve models with NVIDIA NIM as a drop-in replacement for hosted model APIs.
NVIDIA Dynamo
Serve models with NVIDIA Dynamo on the same GPUs that train them.
KServe
Serve models with KServe on Kubernetes clusters that Vantage Compute manages.
Interactive work
JupyterLab
Open JupyterLab notebooks on your own GPU clusters, with the right hardware attached.
Data and storage
Apache Spark
Run Spark on your own clusters, beside notebooks and batch jobs.
VAST, WEKA and DDN
Connect third-party storage from VAST, WEKA and DDN to your clusters.