Accelerate inference with torch.compile caching
Torch.compile caching accelerates model boot times in PyTorch by 2-3 times.
·
1 просмотров
Torch.compile caching accelerates model boot times in PyTorch by 2-3 times.