overview
What are AWS Inferentia2 Instances?
AWS Inferentia2 Instances, or Inf2, are cutting-edge inference accelerators designed specifically for maximizing performance in AI applications. With the support of the Neuron compiler, these instances deliver transformative benefits for organizations leveraging large language models.
- Up to 2.3 petaflops of compute power.
- Supports six data types for flexible optimization.
- First to enable scale-out distributed inference.