refactor_symbol

[clackhan]

No description provided.

[yuanms2]

explicit 是不是只有在一个参数的时候才有意义？

[clackhan]

explicit 是不是只有在一个参数的时候才有意义？

此处explicit的作用防止Symbol在某些非预期的地方隐士转换为bool值导致在编译时放过某些错误，例如下面代码编译器认为是正确的：

  Symbol<DType> dtype = JUST(DType::Get(DataType::kFloat));
  DataType data_type = DataType::kFloat;
  Symbol<cfg::SbpParallel> sbp = JUST(MakeBroadcastSbpParallel());
  if (data_type == dtype) {
    // do nothing
  }
  if (data_type == sbp) {
    // do nothing
  }

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

PyTorch resnet50 time: 142.3ms (= 7116.2ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 128.3ms (= 6416.0ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.11 (= 142.3ms / 128.3ms)

PyTorch resnet50 time: 83.9ms (= 4197.0ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 74.6ms (= 3729.4ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 83.9ms / 74.6ms)

PyTorch resnet50 time: 57.1ms (= 2857.5ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 48.5ms (= 2422.7ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.18 (= 57.1ms / 48.5ms)

PyTorch resnet50 time: 52.1ms (= 2605.8ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 38.7ms (= 1935.0ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.35 (= 52.1ms / 38.7ms)

PyTorch resnet50 time: 43.9ms (= 2197.3ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 40.8ms (= 2039.7ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 1.08 (= 43.9ms / 40.8ms)