Fix linear bug

[Flowingsun007]

No description provided.

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

PyTorch resnet50 time: 140.1ms (= 7003.7ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 128.3ms (= 6413.9ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.09 (= 140.1ms / 128.3ms)

PyTorch resnet50 time: 84.8ms (= 4237.6ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 74.7ms (= 3733.7ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 84.8ms / 74.7ms)

PyTorch resnet50 time: 58.3ms (= 2917.1ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 48.5ms (= 2426.8ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.20 (= 58.3ms / 48.5ms)

PyTorch resnet50 time: 45.3ms (= 2262.9ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 38.8ms (= 1942.5ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.16 (= 45.3ms / 38.8ms)

PyTorch resnet50 time: 39.7ms (= 1983.7ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 37.1ms (= 1856.7ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 1.07 (= 39.7ms / 37.1ms)