The documentation for version 3.0.0 contains two inconsistent descriptions of neuron

[96XinHe]

The first place say : "...When two layers are of the same size or one layer is twice as large as the previous layer, a skip connection is built." (https://docs.deepmodeling.com/projects/deepmd/en/v3.0.0/train/train-input.html#model-standard-descriptor-se-e2-a-neuron)

But another place says: "The neuron specifies the size of the embedding net. From left to right the members denote the sizes of each hidden layer from the input end to the output end, respectively. If the outer layer is twice the size of the inner layer, then the inner layer is copied and concatenated, then a [ResNet architecture] is built between them." (https://docs.deepmodeling.com/projects/deepmd/en/v3.0.0/model/train-se-e2-a.html).

Inconsistent expression of whether or not ResNet connections can be established between two layers of the same size. Maybe one of the descriptions should be revised？

[njzjz]

The first one is correct.

[wanghan-iapcm]

both of them are correct. the first one may mention the concatenation of the input.

[njzjz]

I believe we should copy the equations of the neural networks from the paper to somewhere in the documentation..