📚 The doc issue

The doc of optim.SGD() shows dampening and weight_decay parameter in order as shown below:

Class torch.optim.SGD(..., momentum=0, dampening=0, weight_decay=0, nesterov=False, ...)

But in Parameters section, dampening and weight_decay parameter are explained in opposite order as shown below:

Parameters

• ...
• momentum (float, optional) – momentum factor (default: 0)
• weight_decay (float, optional) – weight decay (L2 penalty) (default: 0) <- Here
• dampening (float, optional) – dampening for momentum (default: 0) <- Here
• nesterov (bool, optional) – enables Nesterov momentum (default: False)
• ...

Suggest a potential alternative/fix

So in Parameters section, dampening and weight_decay parameter should be explained in order as shown below:

Parameters

• ...
• momentum (float, optional) – momentum factor (default: 0)
• dampening (float, optional) – dampening for momentum (default: 0) <- Here
• weight_decay (float, optional) – weight decay (L2 penalty) (default: 0) <- Here
• nesterov (bool, optional) – enables Nesterov momentum (default: False)
• ...

cc @svekars @brycebortree @sekyondaMeta @vincentqb @jbschlosser @albanD @janeyx99 @crcrpar