Add nn.AdaptiveAvgPool1d and nn.AdaptiveAvgPool3d

[terryzhao127]

• Add nn.AdaptiveAvgPool1d and nn.AdaptiveAvgPool3d ops
• Generalize kernel functions for different ops (1d, 2d, 3d)
• Implement CPU version
• Implement GPU version
• Support data with larger size (according to torch.nn.AdaptiveAvgPoolxd)
• Add test case
• Add functional API
• Add docs
• Support different data_format: channels_first and channels_last

Docs:

[terryzhao127]

• Support different data_format: channels_first and channels_last

@BBuf 除了最后一条，均已实现，暂时不支持channels_last格式的输入

[terryzhao127]

上述问题均已解决 @BBuf

[github-actions]

CI failed, removing label automerge

[github-actions]

CI failed, removing label automerge

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

PyTorch resnet50 time: 139.6ms (= 4189.4ms / 30, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 125.8ms (= 3774.8ms / 30, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.11 (= 139.6ms / 125.8ms)

PyTorch resnet50 time: 83.2ms (= 2495.9ms / 30, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 73.4ms (= 2201.9ms / 30, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 83.2ms / 73.4ms)

PyTorch resnet50 time: 56.3ms (= 1687.8ms / 30, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 48.9ms (= 1468.2ms / 30, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.15 (= 56.3ms / 48.9ms)

PyTorch resnet50 time: 49.8ms (= 1494.7ms / 30, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 46.5ms (= 1394.7ms / 30, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.07 (= 49.8ms / 46.5ms)

PyTorch resnet50 time: 39.8ms (= 1194.0ms / 30, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 49.2ms (= 1477.2ms / 30, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 0.81 (= 39.8ms / 49.2ms)