fix repeat 0-dim tensor bug

[BBuf]

fix repeat 0-dim tensor bug

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

OneFlow resnet50 time: 128.3ms (= 6414.1ms / 50, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 142.7ms (= 7134.6ms / 50, input_shape=[16, 3, 224, 224])
Relative speed: 1.11 (= 142.7ms / 128.3ms)

OneFlow resnet50 time: 74.8ms (= 3739.9ms / 50, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 84.9ms (= 4246.8ms / 50, input_shape=[8, 3, 224, 224])
Relative speed: 1.14 (= 84.9ms / 74.8ms)

OneFlow resnet50 time: 47.7ms (= 2385.3ms / 50, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 58.0ms (= 2898.8ms / 50, input_shape=[4, 3, 224, 224])
Relative speed: 1.22 (= 58.0ms / 47.7ms)

OneFlow resnet50 time: 40.3ms (= 2015.4ms / 50, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 44.8ms (= 2237.7ms / 50, input_shape=[2, 3, 224, 224])
Relative speed: 1.11 (= 44.8ms / 40.3ms)

OneFlow resnet50 time: 40.0ms (= 2000.9ms / 50, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 43.4ms (= 2169.7ms / 50, input_shape=[1, 3, 224, 224])
Relative speed: 1.08 (= 43.4ms / 40.0ms)

OneFlow resnet50 time: 142.1ms (= 7104.7ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 158.1ms (= 7904.4ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
Relative speed: 1.11 (= 158.1ms / 142.1ms)

OneFlow resnet50 time: 88.4ms (= 4419.7ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 104.5ms (= 5223.6ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
Relative speed: 1.18 (= 104.5ms / 88.4ms)

OneFlow resnet50 time: 66.7ms (= 3334.8ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.0ms (= 3702.3ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
Relative speed: 1.11 (= 74.0ms / 66.7ms)

OneFlow resnet50 time: 65.9ms (= 3295.5ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 62.8ms (= 3142.1ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
Relative speed: 0.95 (= 62.8ms / 65.9ms)

OneFlow resnet50 time: 58.1ms (= 2903.3ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 63.8ms (= 3189.7ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
Relative speed: 1.10 (= 63.8ms / 58.1ms)

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

OneFlow resnet50 time: 128.2ms (= 6410.9ms / 50, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 140.6ms (= 7029.6ms / 50, input_shape=[16, 3, 224, 224])
Relative speed: 1.10 (= 140.6ms / 128.2ms)

OneFlow resnet50 time: 74.6ms (= 3732.0ms / 50, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 86.3ms (= 4317.0ms / 50, input_shape=[8, 3, 224, 224])
Relative speed: 1.16 (= 86.3ms / 74.6ms)

OneFlow resnet50 time: 48.1ms (= 2405.4ms / 50, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 59.2ms (= 2960.3ms / 50, input_shape=[4, 3, 224, 224])
Relative speed: 1.23 (= 59.2ms / 48.1ms)

OneFlow resnet50 time: 43.9ms (= 2196.5ms / 50, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 45.1ms (= 2254.2ms / 50, input_shape=[2, 3, 224, 224])
Relative speed: 1.03 (= 45.1ms / 43.9ms)

OneFlow resnet50 time: 42.7ms (= 2134.1ms / 50, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 42.3ms (= 2116.7ms / 50, input_shape=[1, 3, 224, 224])
Relative speed: 0.99 (= 42.3ms / 42.7ms)

OneFlow resnet50 time: 144.2ms (= 7209.9ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 160.9ms (= 8044.4ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
Relative speed: 1.12 (= 160.9ms / 144.2ms)

OneFlow resnet50 time: 93.7ms (= 4686.1ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 99.4ms (= 4969.7ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
Relative speed: 1.06 (= 99.4ms / 93.7ms)

OneFlow resnet50 time: 67.2ms (= 3358.5ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 75.9ms (= 3792.8ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
Relative speed: 1.13 (= 75.9ms / 67.2ms)

OneFlow resnet50 time: 66.9ms (= 3346.3ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 64.2ms (= 3208.5ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
Relative speed: 0.96 (= 64.2ms / 66.9ms)

OneFlow resnet50 time: 64.8ms (= 3237.5ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 76.7ms (= 3834.7ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
Relative speed: 1.18 (= 76.7ms / 64.8ms)