fix bug of random_cpu/gpu_placement calc_device

[marigoold]

No description provided.

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 









❌ OneFlow resnet50 time: 140.0ms (= 13999.1ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 163.8ms (= 16376.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.17 (= 163.8ms / 140.0ms)

OneFlow resnet50 time: 85.7ms (= 8570.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 102.7ms (= 10269.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.20 (= 102.7ms / 85.7ms)

OneFlow resnet50 time: 58.2ms (= 11635.8ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 79.6ms (= 15910.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.37 (= 79.6ms / 58.2ms)

OneFlow resnet50 time: 45.2ms (= 9035.3ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 70.1ms (= 14023.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.55 (= 70.1ms / 45.2ms)

OneFlow resnet50 time: 42.0ms (= 8406.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 76.3ms (= 15266.3ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.82 (= 76.3ms / 42.0ms)

[github-actions]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9714/