diff --git a/tutorial/zh/beginner_source/examples_autograd/tf_two_layer_net.py b/tutorial/zh/beginner_source/examples_autograd/tf_two_layer_net.py
index e5e8c3010..29d847f5a 100644
--- a/tutorial/zh/beginner_source/examples_autograd/tf_two_layer_net.py
+++ b/tutorial/zh/beginner_source/examples_autograd/tf_two_layer_net.py
@@ -18,8 +18,8 @@
 
 # 首先我们设置计算图:
 
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 为输入数据和目标数据创建占位符; 
diff --git a/tutorial/zh/beginner_source/examples_autograd/two_layer_net_autograd.py b/tutorial/zh/beginner_source/examples_autograd/two_layer_net_autograd.py
index a510df483..1fcb526b0 100644
--- a/tutorial/zh/beginner_source/examples_autograd/two_layer_net_autograd.py
+++ b/tutorial/zh/beginner_source/examples_autograd/two_layer_net_autograd.py
@@ -20,8 +20,8 @@
 dtype = torch.FloatTensor
 # dtype = torch.cuda.FloatTensor # 取消注释以在GPU上运行
 
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 创建随机张量来保存输入和输出,并将它们包装在变量中.
diff --git a/tutorial/zh/beginner_source/examples_autograd/two_layer_net_custom_function.py b/tutorial/zh/beginner_source/examples_autograd/two_layer_net_custom_function.py
index 8baf3aac8..6ee419546 100644
--- a/tutorial/zh/beginner_source/examples_autograd/two_layer_net_custom_function.py
+++ b/tutorial/zh/beginner_source/examples_autograd/two_layer_net_custom_function.py
@@ -45,8 +45,8 @@ def backward(ctx, grad_output):
 dtype = torch.FloatTensor
 # dtype = torch.cuda.FloatTensor # 取消注释以在GPU上运行
 
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 创建随机张量来保存输入和输出,并将它们包装在变量中.
diff --git a/tutorial/zh/beginner_source/examples_nn/dynamic_net.py b/tutorial/zh/beginner_source/examples_nn/dynamic_net.py
index 1874f0c91..f80265ca4 100644
--- a/tutorial/zh/beginner_source/examples_nn/dynamic_net.py
+++ b/tutorial/zh/beginner_source/examples_nn/dynamic_net.py
@@ -39,8 +39,8 @@ def forward(self, x):
         return y_pred
 
 
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 创建随机张量来保存输入和输出,并将它们包装在变量中.
diff --git a/tutorial/zh/beginner_source/examples_nn/two_layer_net_module.py b/tutorial/zh/beginner_source/examples_nn/two_layer_net_module.py
index 5104e46ac..8138da5d2 100644
--- a/tutorial/zh/beginner_source/examples_nn/two_layer_net_module.py
+++ b/tutorial/zh/beginner_source/examples_nn/two_layer_net_module.py
@@ -30,9 +30,8 @@ def forward(self, x):
         y_pred = self.linear2(h_relu)
         return y_pred
 
-
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 创建随机张量来保存输入和输出,并将它们包装在变量中. 
diff --git a/tutorial/zh/beginner_source/examples_nn/two_layer_net_nn.py b/tutorial/zh/beginner_source/examples_nn/two_layer_net_nn.py
index 49143d3ba..cd907d728 100644
--- a/tutorial/zh/beginner_source/examples_nn/two_layer_net_nn.py
+++ b/tutorial/zh/beginner_source/examples_nn/two_layer_net_nn.py
@@ -12,8 +12,8 @@
 import torch
 from torch.autograd import Variable
 
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 创建随机张量来保存输入和输出,并将它们包装在变量中.
diff --git a/tutorial/zh/beginner_source/examples_nn/two_layer_net_optim.py b/tutorial/zh/beginner_source/examples_nn/two_layer_net_optim.py
index d9480ce8c..ffddff017 100644
--- a/tutorial/zh/beginner_source/examples_nn/two_layer_net_optim.py
+++ b/tutorial/zh/beginner_source/examples_nn/two_layer_net_optim.py
@@ -13,8 +13,8 @@
 import torch
 from torch.autograd import Variable
 
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 创建随机张量来保存输入和输出,并将它们包装在变量中.
diff --git a/tutorial/zh/beginner_source/examples_tensor/two_layer_net_tensor.py b/tutorial/zh/beginner_source/examples_tensor/two_layer_net_tensor.py
index 0b83efad7..b53a2ebdd 100644
--- a/tutorial/zh/beginner_source/examples_tensor/two_layer_net_tensor.py
+++ b/tutorial/zh/beginner_source/examples_tensor/two_layer_net_tensor.py
@@ -18,8 +18,8 @@
 dtype = torch.FloatTensor
 # dtype = torch.cuda.FloatTensor # 取消注释以在GPU上运行
 
-# N 批量大小; D_in是输入尺寸;
-# H是隐藏尺寸; D_out是输出尺寸.
+# N 是一个batch的样本数量; D_in是输入维度;
+# H 是隐藏层向量的维度; D_out是输出维度.
 N, D_in, H, D_out = 64, 1000, 100, 10
 
 # 创建随机输入和输出数据