.to(device) 在 pytorch 中的效率

Question

为了实现最大效率，哪个命令更有效？x = torch.randn(100, 100).to(device)x = torch.randn(100, 100, device = device)在进行繁重的张量运算时，使用一个与另一个相比有什么好处吗？有人告诉我其中一个效率较低，但无法正确计算如何比较两者。我假设第二个更好，因为它直接导出到设备，而不是先创建张量，然后必须将其转移到device.

Answer 1

好吧，我不是这方面的专家。这是给你的计时结果。

%timeit torch.randn(100, 100).to(device)

The slowest run took 12.65 times longer than the fastest. This could mean that an intermediate result is being cached.

10000 loops, best of 3: 129 µs per loop

%timeit torch.randn(100, 100, device = device)

The slowest run took 88.54 times longer than the fastest. This could mean that an intermediate result is being cached.

100000 loops, best of 3: 11.6 µs per loop

PS 我在 Google Colab 上执行了这两个命令。