apply flake8 E203 rule (#684)

tinygrad/nn/__init__.py

@@ -59,8 +59,8 @@ class Linear:
 class GroupNorm:
   def __init__(self, num_groups:int, num_channels:int, eps:float=1e-5, affine:bool=True):
     self.num_groups, self.num_channels, self.eps = num_groups, num_channels, eps
-    self.weight : Optional[Tensor] = Tensor.ones(num_channels) if affine else None
-    self.bias : Optional[Tensor] = Tensor.zeros(num_channels) if affine else None
+    self.weight: Optional[Tensor] = Tensor.ones(num_channels) if affine else None
+    self.bias: Optional[Tensor] = Tensor.zeros(num_channels) if affine else None
 
   def __call__(self, x:Tensor):
     # reshape for layernorm to work as group norm

tinygrad/nn/optim.py

@@ -3,13 +3,13 @@ from typing import List
 from tinygrad.tensor import Tensor
 
 class Optimizer:
-  def __init__(self, params : List[Tensor]):
+  def __init__(self, params: List[Tensor]):
     # if it's None, but being put into an optimizer, set it to True
     for x in params:
       if x.requires_grad is None: x.requires_grad = True
 
-    self.params : List[Tensor] = [x for x in params if x.requires_grad]
-    self.buffers : List[Tensor] = [x for x in params if not x.requires_grad]   # buffers are still realized
+    self.params: List[Tensor] = [x for x in params if x.requires_grad]
+    self.buffers: List[Tensor] = [x for x in params if not x.requires_grad]   # buffers are still realized
 
   # TODO: this probably shouldn't change the gradients, just the ones used by the optimizer
   def clipnorm(self, amount=1):
@@ -27,7 +27,7 @@ class Optimizer:
       p.realize()
 
 class SGD(Optimizer):
-  def __init__(self, params : List[Tensor], lr=0.001, momentum=0, nesterov=False):
+  def __init__(self, params: List[Tensor], lr=0.001, momentum=0, nesterov=False):
     super().__init__(params)
     self.lr, self.momentum, self.nesterov = lr, momentum, nesterov
     self.b = [Tensor.zeros(*t.shape, device=params[0].device, requires_grad=False) for t in self.params] if self.momentum else []
@@ -44,7 +44,7 @@ class SGD(Optimizer):
     self.realize(self.b)
 
 class RMSprop(Optimizer):
-  def __init__(self, params : List[Tensor], lr=0.001, decay=0.9, eps=1e-8):
+  def __init__(self, params: List[Tensor], lr=0.001, decay=0.9, eps=1e-8):
     super().__init__(params)
     self.lr, self.decay, self.eps = lr, decay, eps
 
@@ -58,7 +58,7 @@ class RMSprop(Optimizer):
     self.realize(self.v)
 
 class Adam(Optimizer):
-  def __init__(self, params : List[Tensor], lr=0.001, b1=0.9, b2=0.999, eps=1e-8):
+  def __init__(self, params: List[Tensor], lr=0.001, b1=0.9, b2=0.999, eps=1e-8):
     super().__init__(params)
     # NOTE: self.t is a tensor so Adam can be jitted
     self.lr, self.b1, self.b2, self.eps, self.t = lr, b1, b2, eps, Tensor([0], requires_grad=False).realize()
@@ -77,7 +77,7 @@ class Adam(Optimizer):
     self.realize([self.t] + self.m + self.v)
 
 def get_parameters(obj) -> List[Tensor]:
-  parameters : List[Tensor] = []
+  parameters: List[Tensor] = []
   if isinstance(obj, Tensor):
     parameters.append(obj)
   elif isinstance(obj, (list, tuple)):