Add StreamContext::device_type()/DeviceCtx::device_type()

[liujuncheng]

No description provided.

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

OneFlow resnet50 time: 128.2ms (= 6411.2ms / 50, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 145.0ms (= 7248.1ms / 50, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.13 (= 145.0ms / 128.2ms)

OneFlow resnet50 time: 74.9ms (= 3745.9ms / 50, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 90.8ms (= 4541.8ms / 50, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.21 (= 90.8ms / 74.9ms)

OneFlow resnet50 time: 50.9ms (= 2543.4ms / 50, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 59.2ms (= 2962.4ms / 50, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.16 (= 59.2ms / 50.9ms)

OneFlow resnet50 time: 50.2ms (= 2509.6ms / 50, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 49.0ms (= 2452.5ms / 50, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 0.98 (= 49.0ms / 50.2ms)

OneFlow resnet50 time: 45.5ms (= 2274.1ms / 50, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 42.8ms (= 2141.0ms / 50, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 0.94 (= 42.8ms / 45.5ms)

OneFlow resnet50 time: 157.8ms (= 7889.7ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 163.0ms (= 8149.8ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.03 (= 163.0ms / 157.8ms)

OneFlow resnet50 time: 101.1ms (= 5053.2ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 109.5ms (= 5477.0ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.08 (= 109.5ms / 101.1ms)

OneFlow resnet50 time: 80.3ms (= 4014.6ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.3ms (= 3916.1ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 0.98 (= 78.3ms / 80.3ms)

OneFlow resnet50 time: 71.3ms (= 3566.8ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 69.8ms (= 3491.6ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 0.98 (= 69.8ms / 71.3ms)

OneFlow resnet50 time: 71.5ms (= 3574.4ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 66.9ms (= 3343.1ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 0.94 (= 66.9ms / 71.5ms)