We’re excited to announce that OctoML is now open for Early Access!
OctoML is designed to support machine learning teams looking to improve their model performance and deployments — by increasing model throughput or decreasing latency, reducing inference costs, or simplifying model time-to-market.
As an Early Access partner, you’ll receive:
Free, limited-time access to the OctoML platform.
Personalized, accelerated support from our Engineering and Support teams.
Future discounts on general availability pricing.
OctoML is the SaaS version of the open-source Apache TVM project, and is already in use by some of the largest technology companies in the world. Using a machine learning search function, OctoML optimizes models on a chip-by-chip basis, delivering 50–90% inference performance improvements.
We’ve got answers to all your questions about protecting your model IP, legal terms and conditions, and transitioning to General Access when it becomes available — sign up here and we’ll set up time to chat.