torch.Tensor.register_hook()怎么使用

本篇内容主要讲解"torch.Tensor.register_hook()怎么使用"，感兴趣的朋友不妨来看看。本文介绍的方法操作简单快捷，实用性强。下面就让小编来带大家学习"torch.Tensor.register_hook()怎么使用"吧!

代码实验举例1:

C:\Users\chenxuqi>conda activate ssd4pytorch2_2_0(ssd4pytorch2_2_0) C:\Users\chenxuqi>pythonPython 3.7.7 (default, May  6 2020, 11:45:54) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32Type "help", "copyright", "credits" or "license" for more information.>>> import torch>>> torch.manual_seed(seed=20200910)>>>>>> v = torch.tensor([0., 0., 0.], requires_grad=True)>>> h = v.register_hook(lambda grad: grad * 2)  # double the gradient>>> print(v.grad)None>>> v.backward(torch.tensor([1., 2., 3.]))>>> print(v.grad)tensor([2., 4., 6.])>>> v.gradtensor([2., 4., 6.])>>> h.remove()  # removes the hook>>> v.grad.zero_()tensor([0., 0., 0.])>>> v.gradtensor([0., 0., 0.])>>> v.backward(torch.tensor([1., 2., 3.]))>>> v.gradtensor([1., 2., 3.])>>>>>>

代码实验举例2:

C:\Users\chenxuqi>conda activate ssd4pytorch2_2_0(ssd4pytorch2_2_0) C:\Users\chenxuqi>pythonPython 3.7.7 (default, May  6 2020, 11:45:54) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32Type "help", "copyright", "credits" or "license" for more information.>>> import torch>>> torch.manual_seed(seed=20200910)>>>>>> v = torch.tensor([0., 0., 0.], requires_grad=True)>>> print(v.grad)None>>> v.backward(torch.tensor([1., 2., 3.]))>>> print(v.grad)tensor([1., 2., 3.])>>>>>>>>>

到此，相信大家对"torch.Tensor.register_hook()怎么使用"有了更深的了解，不妨来实际操作一番吧！这里是网站，更多相关内容可以进入相关频道进行查询，关注我们，继续学习！