Fix stack backward

[simonJJJ]

No description provided.

[github-actions]

Static analysis with clang failed. PR label automerge has been removed

[github-actions]

CI failed when running job: cuda-module. PR label automerge has been removed

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

OneFlow resnet50 time: 136.7ms (= 13674.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 139.9ms (= 13988.2ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.02 (= 139.9ms / 136.7ms)

OneFlow resnet50 time: 78.7ms (= 7865.6ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 84.0ms (= 8398.0ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.07 (= 84.0ms / 78.7ms)

OneFlow resnet50 time: 52.2ms (= 10449.0ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 53.5ms (= 10703.9ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.02 (= 53.5ms / 52.2ms)

OneFlow resnet50 time: 43.1ms (= 8622.4ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 44.7ms (= 8943.4ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.04 (= 44.7ms / 43.1ms)

OneFlow resnet50 time: 40.5ms (= 8106.4ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 40.8ms (= 8151.5ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.01 (= 40.8ms / 40.5ms)

OneFlow resnet50 time: 148.8ms (= 14885.0ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 161.2ms (= 16119.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.08 (= 161.2ms / 148.8ms)

OneFlow resnet50 time: 89.8ms (= 8980.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 101.9ms (= 10194.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.14 (= 101.9ms / 89.8ms)

OneFlow resnet50 time: 63.9ms (= 12782.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 72.5ms (= 14491.1ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.13 (= 72.5ms / 63.9ms)

OneFlow resnet50 time: 59.4ms (= 11880.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 61.0ms (= 12205.5ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.03 (= 61.0ms / 59.4ms)

OneFlow resnet50 time: 53.9ms (= 10780.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 58.7ms (= 11737.9ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.09 (= 58.7ms / 53.9ms)