Computer Vision | Vision AI
Vision models,
from cloud to edge.
Train detection, tracking and image-analysis models on Supercloud, then deploy them to cameras and edge devices and follow every version from one console.
Tasks
Detection, tracking and image analysis
Training
On Supercloud clusters
Deployment
Cameras and edge devices
From labelled images to every camera
Train on Supercloud clusters, keep each version in a registry, and deploy it to the devices that run it.
What you get
Training on managed clusters
Train on Kubernetes or Slurm clusters that Supercloud sets up and keeps running.
Versions you can trace
Each model version keeps the data, settings and code it was trained with.
Deployment to the edge
Push a version to cameras and edge devices, and roll it back if it misbehaves.
Every device on the version you chose
See which version each camera and edge device runs, and roll out a new one from the console.
Rollouts in stages
Send a new version to a few devices first, then to the rest.
Health from the edge
Follow latency and errors reported by each device.
Your images stay yours
Training images stay in your project, and on your infrastructure if Supercloud runs there.
FAQ
Which tasks does it cover?
Object detection, tracking and image analysis. Tell us about your use case.
Which devices can run the models?
Tell us your cameras and edge hardware, and we’ll confirm what is supported.
Can we train on our own infrastructure?
Yes. Supercloud can run on your infrastructure, so training images stay there.