Runs
Scale to zeroEphemeral batch compute
Run batch workloads as ephemeral containers that start instantly and terminate on exit. Spawn dozens of parallel runs from your server code, mount shared volumes for feature caches, and get exit codes and logs back — no infrastructure to manage. Pay only for the compute you use.
Overview
Pricing
Usage
Docs
Examples
Parallel Training Runstrain.ts
typescript
import { sylphx } from '@sylphx/sdk'
const folds = [1, 2, 3, 4, 5]
// Spawn all runs in parallel
const runs = await Promise.all(
folds.map(fold =>
sylphx.runs.create({
image: 'registry.sylphx.com/myorg/trainer:abc123',
command: ['python', 'train.py', '--fold', String(fold)],
resources: { requests: { cpu: '4', memory: '8Gi' } },
volumeMounts: [{ volumeId: 'vol_featurecache', mountPath: '/cache' }],
timeoutSeconds: 3600,
})
)
)
// Wait for all to complete
const results = await Promise.all(runs.map(r => sylphx.runs.wait(r.id)))
console.log('All folds done:', results.map(r => r.exitCode))Looking for more examples?
View on GitHub