Set default LR value of SGD to 1e-3

torch/optim/sgd.py

@@ -1,16 +1,17 @@
 import torch
 from torch import Tensor
-from .optimizer import (Optimizer, required, _use_grad_for_differentiable, _default_to_fused_or_foreach,
+from .optimizer import (Optimizer, _use_grad_for_differentiable, _default_to_fused_or_foreach,
                         _differentiable_doc, _foreach_doc, _maximize_doc)
 from typing import List, Optional
 
 __all__ = ['SGD', 'sgd']
 
+
 class SGD(Optimizer):
-    def __init__(self, params, lr=required, momentum=0, dampening=0,
+    def __init__(self, params, lr=1e-3, momentum=0, dampening=0,
                  weight_decay=0, nesterov=False, *, maximize: bool = False, foreach: Optional[bool] = None,
                  differentiable: bool = False):
-        if lr is not required and lr < 0.0:
+        if lr < 0.0:
             raise ValueError(f"Invalid learning rate: {lr}")
         if momentum < 0.0:
             raise ValueError(f"Invalid momentum value: {momentum}")
@@ -51,7 +52,6 @@ def _init_group(self, group, params_with_grad, d_p_list, momentum_buffer_list):
 
         return has_sparse_grad
 
-
     @_use_grad_for_differentiable
     def step(self, closure=None):
         """Performs a single optimization step.
@@ -130,7 +130,7 @@ def step(self, closure=None):
     Args:
         params (iterable): iterable of parameters to optimize or dicts defining
             parameter groups
-        lr (float): learning rate
+        lr (float, optional): learning rate (default: 1e-3)
         momentum (float, optional): momentum factor (default: 0)
         weight_decay (float, optional): weight decay (L2 penalty) (default: 0)
         dampening (float, optional): dampening for momentum (default: 0)

torch/optim/sgd.pyi

@@ -4,7 +4,7 @@ class SGD(Optimizer):
     def __init__(
         self,
         params: ParamsT,
-        lr: float,
+        lr: float = ...,
         momentum: float = ...,
         dampening: float = ...,
         weight_decay: float = ...,