Update RNN training

Section06-Neural-Machine-Translation/section06.tex

@@ -2669,14 +2669,36 @@ $\textrm{``you''} = \argmax_{y} \textrm{P}(y|\textbf{s}_1, \alert{\textbf{C}})$
 %%% 训练
 \begin{frame}{训练 - 整体流程}
     \begin{itemize}
-        \item 有了一个NMT模型，我们应该怎么使用梯度下降算法来训练一个``聪明''的翻译模型呢？
-        \begin{enumerate}[1]
-            \item 参数初始化
-            \item 优化器选择
-            \item 学习率调度
-            \item 多设备加速
-        \end{enumerate}
+        \item 有了一个NMT模型，我们应该怎么使用梯度下降算法来训练一个翻译模型呢？ 或者说哪些因素会对RNN训练产生影响？
     \end{itemize}
+    \begin{beamerboxesrounded}[upper=uppercolblue,lower=lowercolblue,shadow=true]{\small{\textbf{参数初始化}}}
+        {\footnotesize
+        \begin{spacing}{0.9}
+            给定模型结构，初始化的好坏决定了模型最后的性能。
+        \end{spacing}
+        }
+    \end{beamerboxesrounded}
+    \begin{beamerboxesrounded}[upper=uppercolblue,lower=lowercolblue,shadow=true]{\small{\textbf{优化器选择}}}
+        {\footnotesize
+        \begin{spacing}{0.9}
+            选择不同的优化器需要对使用的便利性与效果进行权衡。
+        \end{spacing}
+        }
+    \end{beamerboxesrounded}
+    \begin{beamerboxesrounded}[upper=uppercolblue,lower=lowercolblue,shadow=true]{\small{\textbf{学习率调度}}}
+        {\footnotesize
+        \begin{spacing}{0.9}
+            合适的学习率调度方案可以让训练过程又好又快。
+        \end{spacing}
+        }
+    \end{beamerboxesrounded}
+    \begin{beamerboxesrounded}[upper=uppercolblue,lower=lowercolblue,shadow=true]{\small{\textbf{多设备加速}}}
+        {\footnotesize
+        \begin{spacing}{0.9}
+            当训练非常缓慢的时候，可以使用多个设备并行计算加速。
+        \end{spacing}
+        }
+    \end{beamerboxesrounded}
 \end{frame}
 
 \begin{frame}{训练 - 初始化}
@@ -2737,7 +2759,7 @@ $\textrm{``you''} = \argmax_{y} \textrm{P}(y|\textbf{s}_1, \alert{\textbf{C}})$
                     ylabel style={yshift=-2.5em},xlabel style={yshift=1.5em},
                     legend style={yshift=-6pt, legend plot pos=right,font=\scriptsize,cells={anchor=west}}
                 ]
-                \addplot[orange,line width=1.25pt] coordinates {(0,0) (4,0.7) (5,0.63) (6,0.57) (7,0.525) (8,0.49) (9,0.465) (10,0.44) (11,0.42) (12,0.4)};
+                \addplot[orange,line width=1.25pt] coordinates {(329,0.000045) (447,0.000078) (540,0.00012) (661,0.0002) (752,0.00032) (856,0.00051) (975,0.00089) (996,0.001) (6599,0.001) (6624,0.0005) (7200,0.0005) (7218,0.00025) (7784,0.00025) (7821,0.000125) (8398,0.000125)};
                 \end{axis}
             }
             \end{tikzpicture}