add_tensor_api_doc

[clackhan]

No description provided.

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

OneFlow resnet50 time: 128.2ms (= 6408.2ms / 50, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 141.0ms (= 7051.4ms / 50, input_shape=[16, 3, 224, 224])
Relative speed: 1.10 (= 141.0ms / 128.2ms)

OneFlow resnet50 time: 74.6ms (= 3732.2ms / 50, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 83.4ms (= 4168.9ms / 50, input_shape=[8, 3, 224, 224])
Relative speed: 1.12 (= 83.4ms / 74.6ms)

OneFlow resnet50 time: 48.3ms (= 2415.2ms / 50, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 62.4ms (= 3118.2ms / 50, input_shape=[4, 3, 224, 224])
Relative speed: 1.29 (= 62.4ms / 48.3ms)

OneFlow resnet50 time: 43.0ms (= 2151.5ms / 50, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 44.2ms (= 2210.2ms / 50, input_shape=[2, 3, 224, 224])
Relative speed: 1.03 (= 44.2ms / 43.0ms)

OneFlow resnet50 time: 40.8ms (= 2041.4ms / 50, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 45.1ms (= 2255.7ms / 50, input_shape=[1, 3, 224, 224])
Relative speed: 1.10 (= 45.1ms / 40.8ms)

OneFlow resnet50 time: 141.7ms (= 7086.6ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 162.8ms (= 8138.8ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
Relative speed: 1.15 (= 162.8ms / 141.7ms)

OneFlow resnet50 time: 90.8ms (= 4540.5ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 103.6ms (= 5180.6ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
Relative speed: 1.14 (= 103.6ms / 90.8ms)

OneFlow resnet50 time: 66.7ms (= 3335.6ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 93.6ms (= 4680.0ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
Relative speed: 1.40 (= 93.6ms / 66.7ms)

OneFlow resnet50 time: 62.4ms (= 3122.1ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 67.0ms (= 3351.0ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
Relative speed: 1.07 (= 67.0ms / 62.4ms)

OneFlow resnet50 time: 64.4ms (= 3218.5ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 65.2ms (= 3257.9ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
Relative speed: 1.01 (= 65.2ms / 64.4ms)