Skip to main content
Launch dedicated virtual machines with SSH access for long-running workloads, development environments, or persistent compute needs.

Overview

Lyceum VMs provide full Linux instances with root access, ideal for workloads that need persistent state, custom system configurations, or interactive development sessions.

GPU Support

A100 and H100 GPUs for ML training and inference

SSH Access

Full shell access with your SSH keys

Flexible Resources

Configure CPU, memory, disk, and GPU count

When to Use VMs

  • Use VMs For
  • Consider Alternatives
  • Long-running training jobs (hours/days)
  • Interactive development and debugging
  • Persistent development environments
  • Workloads requiring custom system packages
  • Multi-process applications
  • Jobs requiring specific hardware configurations

Hardware Profiles

ProfileGPUvCPUMemoryBest For
cpuNone416 GBDevelopment, testing, CPU workloads
a100NVIDIA A100864 GBML training, large models
h100NVIDIA H100880 GBLatest ML workloads, maximum performance
Check real-time availability with lyceum vms availability before starting instances.

VM Lifecycle

StatusDescription
pendingVM is being provisioned
ready / runningVM is available for SSH access
failed / errorProvisioning failed
terminatedVM has been deleted

Quick Start

1

Authenticate

lyceum auth login
2

Check availability

lyceum vms availability
3

Start a VM

lyceum vms start -h a100 -k "$(cat ~/.ssh/id_rsa.pub)" -n "my-training-vm"
4

Connect via SSH

The CLI displays the SSH command when your VM is ready:
ssh -i ~/.ssh/id_rsa ubuntu@<ip-address>
5

Terminate when done

lyceum vms terminate <vm_id>

What’s Next