#torch.compile (1)

Accelerate inference with torch.compile caching

Torch.compile caching accelerates model boot times in PyTorch by 2-3 times.