Model distillation

Compressing a large model’s behavior into a smaller one for cost and latency reduction.

Model distillation — Compressing a large model’s behavior into a smaller one for cost and latency reduction..