Torch compile caching for inference speed

Replicate Blog
Generative AI AI Research

Cache your compiled models for faster boot and inference times