add batchnorm3d module

[YongtaoShi]

[daquexian]

NOTE: 这个文件是我改的

[github-actions]

CI failed, removing label automerge

[github-actions]

CI failed, removing label automerge

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

PyTorch resnet50 time: 140.5ms (= 7026.3ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 127.0ms (= 6350.9ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.11 (= 140.5ms / 127.0ms)

PyTorch resnet50 time: 85.0ms (= 4250.6ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 74.2ms (= 3710.7ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.15 (= 85.0ms / 74.2ms)

PyTorch resnet50 time: 57.7ms (= 2884.7ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 47.5ms (= 2376.7ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.21 (= 57.7ms / 47.5ms)

PyTorch resnet50 time: 47.1ms (= 2355.4ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 41.5ms (= 2074.5ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.14 (= 47.1ms / 41.5ms)

PyTorch resnet50 time: 41.5ms (= 2074.4ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 45.6ms (= 2280.1ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 0.91 (= 41.5ms / 45.6ms)