Update 5-2,模型层.ipynb

lyhue1991 · Aug 2, 2024 · ecfd87b · ecfd87b
1 parent 676a73c
commit ecfd87b
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/5-2,模型层.ipynb b/5-2,模型层.ipynb
@@ -53,7 +53,7 @@
     "\n",
     "* nn.Linear：全连接层。参数个数 = 输入层特征数× 输出层特征数(weight)＋ 输出层特征数(bias)\n",
     "\n",
-    "* nn.Embedding：嵌入层。一种比Onehot更加有效的对离散特征进行编码的方法。一般用于将输入中的单词映射为稠密向量。嵌入层的参数需要学习。\n",
+    "* nn.Embedding：嵌入层，一种比Onehot更加有效的对离散特征进行编码的方法。一般用于将输入中的单词映射为稠密向量。嵌入层的参数需要学习。\n",
     "\n",
     "* nn.Flatten：压平层，用于将多维张量样本压成一维张量样本。\n",
     "\n",
@@ -69,7 +69,7 @@
     "\n",
     "* nn.Dropout3d：三维随机丢弃层。\n",
     "\n",
-    "* nn.Threshold：限幅层。当输入大于或小于阈值范围时，截断之。\n",
+    "* nn.Threshold：限幅层，当输入大于或小于阈值范围时，截断之。\n",
     "\n",
     "* nn.ConstantPad2d： 二维常数填充层。对二维张量样本填充常数扩展长度。\n",
     "\n",