Dev scalar op

[MARD1NO]

增加logical scalar系列op

并去除原始实现中不必要的Tensor构造

[Ldpe2G]

下面这些宏，还可以再合并一下，参考

oneflow/oneflow/user/ops/math_binary_broadcast_seq.h

Line 33 in 00b43eb

#define MATH_BINARY_BROADCAST_LOGICAL_FUNC_SEQ \

oneflow/oneflow/user/kernels/math_binary_broadcast_kernels.cpp

Line 68 in 120a7f8

#define REGISTER_MATH_BINARY_BROADCAST_LOGICAL_KERNEL(math_type_pair, device, data_type_pair) \

[MARD1NO]

我觉得用这些宏，观感太不清晰了。

[hjchen2]

后向算子也要支持一下

[MARD1NO]

后向算子也要支持一下

logical算子我们和pytorch都没有后向

[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: 140.6ms (= 7029.9ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 128.1ms (= 6404.7ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.10 (= 140.6ms / 128.1ms)

PyTorch resnet50 time: 84.1ms (= 4206.4ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 74.6ms (= 3727.6ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 84.1ms / 74.6ms)

PyTorch resnet50 time: 58.1ms (= 2906.2ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 48.2ms (= 2411.5ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.21 (= 58.1ms / 48.2ms)

PyTorch resnet50 time: 49.5ms (= 2472.8ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 42.3ms (= 2117.3ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.17 (= 49.5ms / 42.3ms)

PyTorch resnet50 time: 45.4ms (= 2271.1ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 42.4ms (= 2121.8ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 1.07 (= 45.4ms / 42.4ms)

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

PyTorch resnet50 time: 138.1ms (= 6904.0ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 128.1ms (= 6405.6ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.08 (= 138.1ms / 128.1ms)

PyTorch resnet50 time: 84.5ms (= 4224.9ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 74.5ms (= 3726.0ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 84.5ms / 74.5ms)

PyTorch resnet50 time: 56.6ms (= 2829.1ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 47.6ms (= 2380.7ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.19 (= 56.6ms / 47.6ms)

PyTorch resnet50 time: 47.6ms (= 2382.0ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 39.8ms (= 1987.5ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.20 (= 47.6ms / 39.8ms)

PyTorch resnet50 time: 44.4ms (= 2218.3ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 37.9ms (= 1894.0ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 1.17 (= 44.4ms / 37.9ms)

[chengtbf]

这里为什么要都注释掉呢？