medium10 pts · first answer countscost-managementcapacity-planning
Cost per 1,000 predictions
A GPU instance costs $3 per hour and steadily serves 50 predictions per second. What is the cost per 1,000 predictions?
Discussion 0
Join free to post in the discussion.
Related questions
- Throughput of four replicasMLOps Engineer
- Cost of a GPU training jobMLOps Engineer
- Concurrent requests at peakMLOps Engineer
- Training on spot GPUsMLOps Engineer
- A pod stuck in PendingSite Reliability Engineer
- Little's Law for capacitySite Reliability Engineer
- Which pods are evicted first?Site Reliability Engineer
- Contingency vs management reserveProject Manager