RTH_update_op_test

[rentainhe]

本次提交的单测内容如下:

• flow.eq
• flow.tensor.eq
• flow.equal （感觉是没有和torch对齐）
• flow.exp
• flow.tensor.exp
• flow.erf
• flow.tensor.erf
• flow.erfc
• flow.tensor.erfc
• flow.round
• flow.tensor.round
• flow.tensor.expand

Flatten相关的天宇已经完成了

• flow.nn.Flatten
• flow.flatten

To相关的不需要测试

• flow.to
• flow.tensor.to

[BBuf]

看了一下torch.equal，和我们的语意确实没有对齐，torch.equal是判断两个Tensor是不是完全一样，shape和value，你解决一下吧

[rentainhe]

好的，OK

[rentainhe]

这个我到时候单独作为一个PR提交吧

[BBuf]

好，刚试了下还比较麻烦，先记一下TODO吧，下个pr做

[rentainhe]

好的我这边会记着然后去做一下，这个PR可以先Merge

[github-actions]

CI failed, removing label automerge

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

PyTorch resnet50 time: 140.5ms (= 7025.3ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 127.6ms (= 6381.0ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.10 (= 140.5ms / 127.6ms)

PyTorch resnet50 time: 83.8ms (= 4191.5ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 74.2ms (= 3709.3ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 83.8ms / 74.2ms)

PyTorch resnet50 time: 56.0ms (= 2801.6ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 49.7ms (= 2484.1ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 56.0ms / 49.7ms)

PyTorch resnet50 time: 44.7ms (= 2234.2ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 40.7ms (= 2034.7ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.10 (= 44.7ms / 40.7ms)

PyTorch resnet50 time: 43.4ms (= 2167.8ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 46.3ms (= 2314.9ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 0.94 (= 43.4ms / 46.3ms)