[CodeStyle][Typos][S-[43-49]] Fix typo (storage, sotring...)

_typos.toml

@@ -99,13 +99,6 @@ Stati = 'Stati'
 STOPED = 'STOPED'
 Stoped = 'Stoped'
 stoped = 'stoped'
-storeage = 'storeage'
-sotring = 'sotring'
-stragety = 'stragety'
-strem = 'strem'
-structed = 'structed'
-sturcture = 'sturcture'
-subsituted = 'subsituted'
 subsitute = 'subsitute'
 substitude = 'substitude'
 substitue = 'substitue'

paddle/fluid/framework/new_executor/interpreter/interpreter_util.cc

@@ -61,7 +61,7 @@ namespace paddle::framework::interpreter {
 
 using VariableIdMap = std::map<std::string, std::vector<int>>;
 
-// NOTE(Ruibiao): SingleStreamGuard make some multi-strem op (i.e.,
+// NOTE(Ruibiao): SingleStreamGuard make some multi-stream op (i.e.,
 // c_allreduce_sum) run in single stream. It is dedicated to BuildOpFuncList
 // which run kernel without stream synchronization.
 class SingleStreamGuard {

paddle/phi/kernels/gpu/batch_norm_grad_kernel.cu

@@ -175,7 +175,7 @@ static __global__ LAUNCH_BOUNDS(BlockDim) void BNBackward(
   __shared__ typename BlockReduce::TempStorage ds_storage;
   __shared__ typename BlockReduce::TempStorage db_storage;
   __shared__ typename BlockReduce::TempStorage mean_storage;
-  __shared__ typename BlockReduce::TempStorage variance_storeage;
+  __shared__ typename BlockReduce::TempStorage variance_storage;
   __shared__ BatchNormParamType<T> inv_var_val;
   __shared__ BatchNormParamType<T> mean_val;
   __shared__ BatchNormParamType<T> dscale_val;
@@ -207,7 +207,7 @@ static __global__ LAUNCH_BOUNDS(BlockDim) void BNBackward(
 
       x_sum = BlockReduce(mean_storage).Reduce(x_sum, cub::Sum());
       x_square_sum =
-          BlockReduce(variance_storeage).Reduce(x_square_sum, cub::Sum());
+          BlockReduce(variance_storage).Reduce(x_square_sum, cub::Sum());
       if (threadIdx.x == 0) {
         mean_val = x_sum / inner_size;
         inv_var_val =

paddle/phi/kernels/gpu/batch_norm_kernel.cu

@@ -130,7 +130,7 @@ static __global__ LAUNCH_BOUNDS(BlockDim) void BNForwardTraining(
   int inner_size = N * HxW;
   typedef cub::BlockReduce<BatchNormParamType<T>, BlockDim> BlockReduce;
   __shared__ typename BlockReduce::TempStorage mean_storage;
-  __shared__ typename BlockReduce::TempStorage variance_storeage;
+  __shared__ typename BlockReduce::TempStorage variance_storage;
   __shared__ BatchNormParamType<T> mean_val;
   __shared__ BatchNormParamType<T> variance_val;
   __shared__ BatchNormParamType<T> inv_var_val;
@@ -149,7 +149,7 @@ static __global__ LAUNCH_BOUNDS(BlockDim) void BNForwardTraining(
     }
     x_sum = BlockReduce(mean_storage).Reduce(x_sum, cub::Sum());
     x_square_sum =
-        BlockReduce(variance_storeage).Reduce(x_square_sum, cub::Sum());
+        BlockReduce(variance_storage).Reduce(x_square_sum, cub::Sum());
     if (threadIdx.x == 0) {
       mean_val = x_sum / inner_size;
       variance_val = x_square_sum / inner_size - mean_val * mean_val;

python/paddle/distributed/auto_parallel/static/helper.py

@@ -463,9 +463,9 @@ def init(self, main_program, place, dist_context):
         if self.lazy_init:
             return
 
-        amp_stragety = dist_context.strategy.amp
-        amp_config = copy.deepcopy(amp_stragety.to_dict())
-        need_cast_parameter = amp_stragety.enable and amp_config["level"] in [
+        amp_strategy = dist_context.strategy.amp
+        amp_config = copy.deepcopy(amp_strategy.to_dict())
+        need_cast_parameter = amp_strategy.enable and amp_config["level"] in [
             "o2",
             "o3",
         ]

python/paddle/nn/functional/loss.py

@@ -4478,7 +4478,7 @@ def adaptive_log_softmax_with_loss(
         name (str|None, optional): Name for the operation (optional, default is ``None``). For more information, please refer to :ref:`api_guide_Name`.
 
     Returns:
-        - output (Tensor). The tensor sotring adaptive logsoftmax result, the shape of output is ``[N]``
+        - output (Tensor). The tensor storing adaptive logsoftmax result, the shape of output is ``[N]``
         - loss (Tensor). The tensor variable storing the adaptive_log_softmax_loss of input and label.
 
     Examples:

test/ir/pir/cinn/test_anchor_fusion.py

@@ -216,7 +216,7 @@ def init():
 
         self.check_accuracy_and_kernel_num(init, func, kernel_num=1)
 
-    def test_recompute_multidownstrema_trivial(self):
+    def test_recompute_multidownstream_trivial(self):
         #     T
         #    / \
         #   S   S

test/legacy_test/test_hsigmoid_op.py

@@ -282,7 +282,7 @@ def test_check_output(self):
 
 
 @skip_check_grad_ci(
-    reason="[skip shape check] The huffman tree is structed separately. It will be complicated if use large shape."
+    reason="[skip shape check] The huffman tree is structured separately. It will be complicated if use large shape."
 )
 class TestHSigmoidOpWithCustomTree(OpTest):
     def setUp(self):
@@ -343,7 +343,7 @@ def test_check_grad(self):
 
 
 @skip_check_grad_ci(
-    reason="[skip shape check] The huffman tree is structed separately. It will be complicated if use large shape."
+    reason="[skip shape check] The huffman tree is structured separately. It will be complicated if use large shape."
 )
 class TestHSigmoidOpWithCustomTreeWithoutBias(OpTest):
     def setUp(self):