From 9486db87d8885bf0bfaee20866c880c4630eb97d Mon Sep 17 00:00:00 2001
From: qiaolongfei <qiaolongfei@baidu.com>
Date: Mon, 27 Mar 2017 10:13:30 +0800
Subject: [PATCH] add draw line in 01.fit-a-line

---
 01.fit_a_line/README.en.ipynb          |  48 ++++++++++++++++++------
 01.fit_a_line/README.en.md             |  46 +++++++++++++++++------
 01.fit_a_line/README.ipynb             |  49 +++++++++++++++++++------
 01.fit_a_line/README.md                |  45 ++++++++++++++++++-----
 01.fit_a_line/image/train-and-test.png | Bin 0 -> 17970 bytes
 01.fit_a_line/index.en.html            |  46 +++++++++++++++++------
 01.fit_a_line/index.html               |  45 ++++++++++++++++++-----
 7 files changed, 213 insertions(+), 66 deletions(-)
 create mode 100644 01.fit_a_line/image/train-and-test.png

diff --git a/01.fit_a_line/README.en.ipynb b/01.fit_a_line/README.en.ipynb
index a819dc5..8330e57 100644
--- a/01.fit_a_line/README.en.ipynb
+++ b/01.fit_a_line/README.en.ipynb
@@ -7,7 +7,7 @@
         "# Linear Regression\n",
         "Let us begin the tutorial with a classical problem called Linear Regression \\[[1](#References)\\]. In this chapter, we will train a model from a realistic dataset to predict home prices. Some important concepts in Machine Learning will be covered through this example.\n",
         "\n",
-        "The source code for this tutorial lives on [book/fit_a_line](https://github.com/PaddlePaddle/book/tree/develop/fit_a_line). For instructions on getting started with PaddlePaddle, see [PaddlePaddle installation guide](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/getstarted/build_and_install/docker_install_en.rst).\n",
+        "The source code for this tutorial lives on [book/fit_a_line](https://github.com/PaddlePaddle/book/tree/develop/fit_a_line). For instructions on getting started with PaddlePaddle, see [PaddlePaddle installation guide](https://github.com/PaddlePaddle/book/blob/develop/README.en.md).\n",
         "\n",
         "## Problem Setup\n",
         "Suppose we have a dataset of $n$ real estate properties. These real estate properties will be referred to as *homes* in this chapter for clarity.\n",
@@ -308,19 +308,41 @@
         "editable": true
       },
       "source": [
-        "# event_handler to print training and testing info\n",
+        "import matplotlib.pyplot as plt\n",
+        "from IPython import display\n",
+        "import cPickle\n",
+        "\n",
+        "step=0\n",
+        "\n",
+        "train_costs=[],[]\n",
+        "test_costs=[],[]\n",
+        "\n",
         "def event_handler(event):\n",
+        "    global step\n",
+        "    global train_costs\n",
+        "    global test_costs\n",
         "    if isinstance(event, paddle.event.EndIteration):\n",
-        "        if event.batch_id % 100 == 0:\n",
-        "            print \"Pass %d, Batch %d, Cost %f\" % (\n",
-        "                event.pass_id, event.batch_id, event.cost)\n",
-        "\n",
-        "    if isinstance(event, paddle.event.EndPass):\n",
-        "        result = trainer.test(\n",
-        "            reader=paddle.batch(\n",
-        "                uci_housing.test(), batch_size=2),\n",
-        "            feeding=feeding)\n",
-        "        print \"Test %d, Cost %f\" % (event.pass_id, result.cost)\n"
+        "        need_plot = False\n",
+        "        if step % 10 == 0:  # every 10 batches, record a train cost\n",
+        "            train_costs[0].append(step)\n",
+        "            train_costs[1].append(event.cost)\n",
+        "\n",
+        "        if step % 1000 == 0: # every 1000 batches, record a test cost\n",
+        "            result = trainer.test(\n",
+        "                reader=paddle.batch(\n",
+        "                    uci_housing.test(), batch_size=2),\n",
+        "                feeding=feeding)\n",
+        "            test_costs[0].append(step)\n",
+        "            test_costs[1].append(result.cost)\n",
+        "\n",
+        "        if step % 100 == 0: # every 100 batches, update cost plot\n",
+        "            plt.plot(*train_costs)\n",
+        "            plt.plot(*test_costs)\n",
+        "            plt.legend(['Train Cost', 'Test Cost'], loc='upper left')\n",
+        "            display.clear_output(wait=True)\n",
+        "            display.display(plt.gcf())\n",
+        "            plt.gcf().clear()\n",
+        "        step += 1\n"
       ],
       "outputs": [
         {
@@ -372,6 +394,8 @@
       "cell_type": "markdown",
       "metadata": {},
       "source": [
+        "\n",
+        "![png](./image/train-and-test.png)\n",
         "\n",
         "## Summary\n",
         "This chapter introduces *Linear Regression* and how to train and test this model with PaddlePaddle, using the UCI Housing Data Set. Because a large number of more complex models and techniques are derived from linear regression, it is important to understand its underlying theory and limitation.\n",
diff --git a/01.fit_a_line/README.en.md b/01.fit_a_line/README.en.md
index 6d06d6a..4d4d0a2 100644
--- a/01.fit_a_line/README.en.md
+++ b/01.fit_a_line/README.en.md
@@ -163,19 +163,41 @@ feeding={'x': 0, 'y': 1}
 Moreover, an event handler is provided to print the training progress:
 
 ```python
-# event_handler to print training and testing info
+import matplotlib.pyplot as plt
+from IPython import display
+import cPickle
+
+step=0
+
+train_costs=[],[]
+test_costs=[],[]
+
 def event_handler(event):
+    global step
+    global train_costs
+    global test_costs
     if isinstance(event, paddle.event.EndIteration):
-        if event.batch_id % 100 == 0:
-            print "Pass %d, Batch %d, Cost %f" % (
-                event.pass_id, event.batch_id, event.cost)
-
-    if isinstance(event, paddle.event.EndPass):
-        result = trainer.test(
-            reader=paddle.batch(
-                uci_housing.test(), batch_size=2),
-            feeding=feeding)
-        print "Test %d, Cost %f" % (event.pass_id, result.cost)
+        need_plot = False
+        if step % 10 == 0:  # every 10 batches, record a train cost
+            train_costs[0].append(step)
+            train_costs[1].append(event.cost)
+
+        if step % 1000 == 0: # every 1000 batches, record a test cost
+            result = trainer.test(
+                reader=paddle.batch(
+                    uci_housing.test(), batch_size=2),
+                feeding=feeding)
+            test_costs[0].append(step)
+            test_costs[1].append(result.cost)
+
+        if step % 100 == 0: # every 100 batches, update cost plot
+            plt.plot(*train_costs)
+            plt.plot(*test_costs)
+            plt.legend(['Train Cost', 'Test Cost'], loc='upper left')
+            display.clear_output(wait=True)
+            display.display(plt.gcf())
+            plt.gcf().clear()
+        step += 1
 ```
 
 ### Start Training
@@ -191,6 +213,8 @@ trainer.train(
     num_passes=30)
 ```
 
+![png](./image/train-and-test.png)
+
 ## Summary
 This chapter introduces *Linear Regression* and how to train and test this model with PaddlePaddle, using the UCI Housing Data Set. Because a large number of more complex models and techniques are derived from linear regression, it is important to understand its underlying theory and limitation.
 
diff --git a/01.fit_a_line/README.ipynb b/01.fit_a_line/README.ipynb
index 07f15fa..38b23e6 100644
--- a/01.fit_a_line/README.ipynb
+++ b/01.fit_a_line/README.ipynb
@@ -7,7 +7,7 @@
         "# 线性回归\n",
         "让我们从经典的线性回归（Linear Regression \\[[1](#参考文献)\\]）模型开始这份教程。在这一章里，你将使用真实的数据集建立起一个房价预测模型，并且了解到机器学习中的若干重要概念。\n",
         "\n",
-        "本教程源代码目录在[book/fit_a_line](https://github.com/PaddlePaddle/book/tree/develop/fit_a_line)， 初次使用请参考PaddlePaddle[安装教程](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/getstarted/build_and_install/docker_install_cn.rst)。\n",
+        "本教程源代码目录在[book/fit_a_line](https://github.com/PaddlePaddle/book/tree/develop/fit_a_line)， 初次使用请参考PaddlePaddle[安装教程](https://github.com/PaddlePaddle/book/blob/develop/README.md)。\n",
         "\n",
         "## 背景介绍\n",
         "给定一个大小为$n$的数据集  ${\\{y_{i}, x_{i1}, ..., x_{id}\\}}_{i=1}^{n}$，其中$x_{i1}, \\ldots, x_{id}$是第$i$个样本$d$个属性上的取值，$y_i$是该样本待预测的目标。线性回归模型假设目标$y_i$可以被属性间的线性组合描述，即\n",
@@ -35,7 +35,7 @@
         "\n",
         "$\\hat{Y}$ 表示模型的预测结果，用来和真实值$Y$区分。模型要学习的参数即：$\\omega_1, \\ldots, \\omega_{13}, b$。\n",
         "\n",
-        "建立模型后，我们需要给模型一个优化目标，使得学到的参数能够让预测值$\\hat{Y}$尽可能地接近真实值$Y$。这里我们引入损失函数（[Loss Function](https://en.wikipedia.org/wiki/Loss_function)，或Cost Function）这个概念。 输入任意一个数据样本的目标值$y_{i}$和模型给出的预测值$\\hat{y_{i}}$，损失函数输出一个非负的实值。这个实质通常用来反映模型误差的大小。\n",
+        "建立模型后，我们需要给模型一个优化目标，使得学到的参数能够让预测值$\\hat{Y}$尽可能地接近真实值$Y$。这里我们引入损失函数（[Loss Function](https://en.wikipedia.org/wiki/Loss_function)，或Cost Function）这个概念。 输入任意一个数据样本的目标值$y_{i}$和模型给出的预测值$\\hat{y_{i}}$，损失函数输出一个非负的实值。这个实值通常用来反映模型误差的大小。\n",
         "\n",
         "对于线性回归模型来讲，最常见的损失函数就是均方误差（Mean Squared Error， [MSE](https://en.wikipedia.org/wiki/Mean_squared_error)）了，它的形式是：\n",
         "\n",
@@ -304,18 +304,41 @@
       },
       "source": [
         "# event_handler to print training and testing info\n",
+        "import matplotlib.pyplot as plt\n",
+        "from IPython import display\n",
+        "import cPickle\n",
+        "\n",
+        "step=0\n",
+        "\n",
+        "train_costs=[],[]\n",
+        "test_costs=[],[]\n",
+        "\n",
         "def event_handler(event):\n",
+        "    global step\n",
+        "    global train_costs\n",
+        "    global test_costs\n",
         "    if isinstance(event, paddle.event.EndIteration):\n",
-        "        if event.batch_id % 100 == 0:\n",
-        "            print \"Pass %d, Batch %d, Cost %f\" % (\n",
-        "                event.pass_id, event.batch_id, event.cost)\n",
-        "\n",
-        "    if isinstance(event, paddle.event.EndPass):\n",
-        "        result = trainer.test(\n",
-        "            reader=paddle.batch(\n",
-        "                uci_housing.test(), batch_size=2),\n",
-        "            feeding=feeding)\n",
-        "        print \"Test %d, Cost %f\" % (event.pass_id, result.cost)\n"
+        "        need_plot = False\n",
+        "        if step % 10 == 0:  # every 10 batches, record a train cost\n",
+        "            train_costs[0].append(step)\n",
+        "            train_costs[1].append(event.cost)\n",
+        "\n",
+        "        if step % 1000 == 0: # every 1000 batches, record a test cost\n",
+        "            result = trainer.test(\n",
+        "                reader=paddle.batch(\n",
+        "                    uci_housing.test(), batch_size=2),\n",
+        "                feeding=feeding)\n",
+        "            test_costs[0].append(step)\n",
+        "            test_costs[1].append(result.cost)\n",
+        "\n",
+        "        if step % 100 == 0: # every 100 batches, update cost plot\n",
+        "            plt.plot(*train_costs)\n",
+        "            plt.plot(*test_costs)\n",
+        "            plt.legend(['Train Cost', 'Test Cost'], loc='upper left')\n",
+        "            display.clear_output(wait=True)\n",
+        "            display.display(plt.gcf())\n",
+        "            plt.gcf().clear()\n",
+        "        step += 1\n"
       ],
       "outputs": [
         {
@@ -367,6 +390,8 @@
       "cell_type": "markdown",
       "metadata": {},
       "source": [
+        "\n",
+        "![png](./image/train-and-test.png)\n",
         "\n",
         "## 总结\n",
         "在这章里，我们借助波士顿房价这一数据集，介绍了线性回归模型的基本概念，以及如何使用PaddlePaddle实现训练和测试的过程。很多的模型和技巧都是从简单的线性回归模型演化而来，因此弄清楚线性模型的原理和局限非常重要。\n",
diff --git a/01.fit_a_line/README.md b/01.fit_a_line/README.md
index a7967db..56fecf7 100644
--- a/01.fit_a_line/README.md
+++ b/01.fit_a_line/README.md
@@ -159,18 +159,41 @@ feeding={'x': 0, 'y': 1}
 
 ```python
 # event_handler to print training and testing info
+import matplotlib.pyplot as plt
+from IPython import display
+import cPickle
+
+step=0
+
+train_costs=[],[]
+test_costs=[],[]
+
 def event_handler(event):
+    global step
+    global train_costs
+    global test_costs
     if isinstance(event, paddle.event.EndIteration):
-        if event.batch_id % 100 == 0:
-            print "Pass %d, Batch %d, Cost %f" % (
-                event.pass_id, event.batch_id, event.cost)
-
-    if isinstance(event, paddle.event.EndPass):
-        result = trainer.test(
-            reader=paddle.batch(
-                uci_housing.test(), batch_size=2),
-            feeding=feeding)
-        print "Test %d, Cost %f" % (event.pass_id, result.cost)
+        need_plot = False
+        if step % 10 == 0:  # every 10 batches, record a train cost
+            train_costs[0].append(step)
+            train_costs[1].append(event.cost)
+
+        if step % 1000 == 0: # every 1000 batches, record a test cost
+            result = trainer.test(
+                reader=paddle.batch(
+                    uci_housing.test(), batch_size=2),
+                feeding=feeding)
+            test_costs[0].append(step)
+            test_costs[1].append(result.cost)
+
+        if step % 100 == 0: # every 100 batches, update cost plot
+            plt.plot(*train_costs)
+            plt.plot(*test_costs)
+            plt.legend(['Train Cost', 'Test Cost'], loc='upper left')
+            display.clear_output(wait=True)
+            display.display(plt.gcf())
+            plt.gcf().clear()
+        step += 1
 ```
 
 ### 开始训练
@@ -186,6 +209,8 @@ trainer.train(
     num_passes=30)
 ```
 
+![png](./image/train-and-test.png)
+
 ## 总结
 在这章里，我们借助波士顿房价这一数据集，介绍了线性回归模型的基本概念，以及如何使用PaddlePaddle实现训练和测试的过程。很多的模型和技巧都是从简单的线性回归模型演化而来，因此弄清楚线性模型的原理和局限非常重要。
 
diff --git a/01.fit_a_line/image/train-and-test.png b/01.fit_a_line/image/train-and-test.png
new file mode 100644
index 0000000000000000000000000000000000000000..bcd304a6a0baf30ecfbc43e08fc0aca179d05958
GIT binary patch
literal 17970
zcmagG1yo!?^Dl@qfnXuQonXO&YjF1f!7afV+}$mBa0%`*xVr>*cXxNUy?pQgz4La@
z?pZi<?wy&Q>h7xQs$W%C-`@&y;wVUXNKjBvD3TH)icnC{zQ79>gaCZPi3NKFT;S}4
zB$YwHpF7AP7<m8AT0-3p3JUqd+Y5SP+4BhsiUdkhL{QoJ?{TWr-;dkRy{n@-uo1Sb
z`xD*1D0FJSerxOZs*&SivYE*UW%09#G2(=aVM>REz>A3d3-l#KkD|O5Xtd{y&?&X{
zxV?X_sB)@m0{)L{Jyy$RTP$~yjE!1HkA^N?(oQF1XYd5k;YjUlJ}*GM-BbjlgM3L0
zILY9ANkF&}ARiKtP$0Pw3CK5U{eQkPWrUL|>=ao<N1mV{Br%l~`_JX~9^XD+gaA!$
z&CHi^M-l199N5qF{SyDV{E50@gL$L=KV;=5ngU}L`z<m6;?cimA+dy3!oLJ7L*sz=
zm6@^6dGIzU@BY6=kyC;`H8piM+x8MSlFGAb&8Z;OuN(bjXlQ72(NCqMrS*4jRs<am
zKS!jDl(w|_LMoXJ+_ThR$Mtm9)juLPpOZsaQ&YoxyJV;3aXn3^RsVH-2jq+H8;c{Z
zXQ(eHhYC<)wOZkYR(Y3M0D<JXT<&nDj)<g?fpC8-4%~41#(sh!|MRX~qCE6T%r}a&
z6T>&wf*c)A5dD`QNDyhwp4)W3l0vi2Qns4Lmn4Uj49+IQa4;Ul3pL1ZXE=r9C=#E8
z;ph<P<@@*V9qU2FM1Q`qv9a;q%;{ep&PF~xxw~E;bX$2|rWE)aK0jRlMEQvgN9uw>
z8VvjC(<c<KAYxC>7%F+oxwiYAl;@4v+1cJ<EC}1ImK}<&$Mq*nN|}u6>eyWA1luK>
z-t%Rr)$Zd~U>vAqaPQW}#fi5c4$F3X@U3!w6st5po-_@ma72;{2P!Bi2qZ9Q*T23z
z78V!7E(vAE{<_>5hFY>6q+jxSeb%z=r>HiY4w;nW371V_S6?|yv1<MKaJBFLu%G{)
zh&S=*$Q%g?>FxIo$F+-4)C>$e*ll<4#|t&tUav1<cx*-MPsE;gK7xJ-<BbkrZij<^
zuxPwc?dry<@T@bQIVlAsISi(g_b;b0s+C1KbSJl5jbJcQ+f!me8H>$^2;a+{eor{Q
zxU4J+txCySH@2?9*;<GBa${U5V{0OL1g9@BnbC`<+ZDISvbLAx%}D+<P0NOWw&!E3
zw|RP==gFlGnC7R7)mbjXEZ9C@&$5_|k?=k2k*h#7x&yFvBvxBJxE_zH?w?P+MuD-l
zhjXk*xVUf?t3Y7Z9#7lyfq5t^D~mkS!3j#xkp>9Rup!;SmurH*je)?3GElD2DL%Sp
z5H#w#$ve?C#f%U2vf!3~oW?k31vc|uCXn8L%%sIq1Bq>>UUZDOsw2FYG7N1`wou6&
zRwMmX=}Xon^67j_F1x8if4NvP(BBhs$HvHUzw-tdog_}=@{f6?)#Gt*B3D{~huvb~
zpD#S}-OZwv=}ZZ(9|Fn-aBF-tSMk^5W`1jHtK4##`{sC2{PE_P>v~FwT(*(~BpH&5
z*Rj|>*v^wz9^`GRahM;ch=%+4il_y1#qlBsuLCc~=D%sP_E`WnG)^(5dIs@VTvT>b
zZNzkL@)qAW*Q>Jy-6|s#ub0bo^=4<LxVX5f*;&M4dmg8grSj!0(NJ%SOQ26BW#x}&
z+!PcP_IqPKPft&Q<!RsCDt~n&L5W2Y4Fh5s@rH)oo6JWG2nf(}-OsI^pauz<p(}!_
zXI5!*UGL#^-~H7$kAyZ_!}6u|aSFrG<EPF=idEDOt5>R?d4!HtRPqk>IscyDPJ(rk
zocFJ;cUaqLxjTwfIPvRG7<$dxt%lvS&C5~#$OpdNbT7WL<_n~^(Ht*KpE}Q$YR@-0
zF$SR$nxXqaH0#*wqY>WYv;)$O9NgaMLgU|B0GDmI`@l}<OQ25#OW=p{m_nSOa9+_^
zI-8&uWzIvw%LUeb1p&3NSg0oD=TDz5R%Nl>>PytR%?f4MSh5@4J}55BD<}ZvPfDt*
z<G%H--xHSUbh`4_enOJ>qK_P~Ugh@tiU2p3$C^w5nH?=RVFz^yrm$NmNUd9Qavqg;
zDM(e=Z1yD1(SU?Df*s08i+^RIemL=h&P9q-T)4zmI*d;AdVO3C!(|BswAW@<+sOt{
z=-#bHDH^`}Elh^yGoe6KLaD<U2p3>i)En(-_+Rfwq2D+Je3osuJMyMp`Dv~pcB9|0
zy&2n{qO&8oLjYRm32jdS?=UH&Qc|P=&3C?A4|>xe8ft32@hnlM3RQIfxt}A_af$)Z
zoAXJ?J=Q25mS%K1$bYkIR({rd8ks;4sdL5f7B`2_dA&3ao5HNioHIX=*`ef8Uf{nI
z+As-e2&9DW3OoMUZk%_f8J9vTP`J+hfKabMGp#~1ja5Ge!USyThp^O>>`j}5Js#@Z
z5S!Z<7rZ+G8VAJkPxT|gFBB)G*@rex#4UsZ;1*q)Yq&LB2iDF>YdXMnea1=()Ck7^
zPhU}{1@M5|S^WR^X@HR?nUVhCio<LWyqyedR8(B-ay2f#zgWizXjG1Bna<NqU7O)z
zt%cwyau489`bOkZa%tfN&jkHRQ2m>f4NJx}?52OHG8qnh<a;_3mXemH`SN8WQ9t5m
zPZ+LEQyB?}CSr!W@O6pu0h5|{es8`iKaJTiL{-hhqlJj+m;cFfQ=R2%YfI~XUdojZ
zEMoPHhN`mi-eFmrPMyV~N}cJXSi0>1jqB^nv+Ld2`og;PUUM^VYI=HE5E}7`Q(1ou
z)gczW#vUP$BV4udU_AXpv^2xG#bWLB^W*KTRok;;SH*ml@gb%hr(m7gw3O4qlqCPv
zn6L&I!sB>U0R2Yk^78UJnnKz<_uCIoazP34u+vc$z7)Re=Q1MS)WJ=!^pW093VCsd
z-^S@4@c_wqfZ$}QHqS}xqvgw0K0lM|mC1_xrR>4s;e4wH4;L5L$4{UBq@)bCy*^v*
z3?+4W+%EUh*Dd{WSuRw{w*`QRE4{|%pu?<o?|y#z$iLs1>aFfv*gDRnfZy8;V;Na%
zeLUIxBTls5cvyO=8E?DR?o&yZOOivgPY%bE&=o)byS5XHTs&e6aB~)F)H~<B#Jg`S
zpDOtNW(18#iP+8O+JWWV5?}saw7`&N7NXf(X*ih3q>td@;_|j4-{Z4~Rp|HH_TU;v
z#l&>U^FMzcR74UA{KF?Sc0U(eA32i@Fwtf|H^CKhv2EKhS7$}+v}A(_SUrQiu}r`r
zSRX>_F8*Q4N5#c`x>z_~tTVnon1+r_h>abq&=&`+`wIF@m9c!DYzjg<&@nX~-BvJp
zM22pgXBZBXZzw|}b;uNwZxm8aB)Z6csp*$kMf<~<aCyFa{|-My%^qB)UvT7NVS1ZA
z;ac|Nq6BtBOiS*U!=wp8ysuA}E|<d`^EF`XRBne0o|PJ~>dn!-@Q)uq{BRNza(NF5
zvfq}xTz<^$s#iL=v_ULB_oM%N0ap|X@21n%lprG4f=uaMK=+b(U6|KyF3U7(kP-$(
z0UOajGGbYPVt+72vo{fypzVlj{MQCSRomf{bOJ;7zP+M`#=szB+t6Hv{{CzkaWA10
zK_uS;H2`Gj)GPg&^m|yZ3|x0po!;1`uCDGmsK}ivS8KV9w@XD15)vw3Ec#o(KT$n9
zHYv{!dAvT%*Q_-!1iV!-GBR?_j7Hn@&0-_Kp+?bAtVX4Lx&NsAzP%ZwCXGoy@MN*B
zjDebZ42}29d;7VlsOXA3k~hok_5`9?ZIbDBb5wZRc*vlwrBzZh<2soq`>+wpI0Fw4
z|3>ivQnzBpc?&4iBp|9;V0BNyBBD;|x}S@7RUi_(D>&_q3BIj1>`$NOrQC|%FkS?@
zSQ)8$q`-g8$~6eEQ2*~{C8km+Dm7IWkU-WjI*fv*W@y)+%CZC!gO4<ixoJ)-E=$&3
zDCTqJJ`WEMA}QpF0K(h`{FVM#hTv~@(Kpy?emkC=j3*X`(^2p}d#=TegCT(ydgh%9
zCjlw~3f@0pk?gOuxMdAv_=kP|dy-8<3pYM?n910DGyBcu9hW+dCf5hRED85u_)A8E
zGXX@cg+U6%ng!r_1BU;idgu+j4{)Sa7^DDf?X%AM0FyhF9;-+jr$`&CxEZH79IMz8
zr+68sco?gQ8LQYJw#2LN)ne1EO7{T>1V}{0{|RNp9w02C{mrQ%0vSMaK#f9;f+3Hv
zh~7M5N+(dLCXh$yQT+30m^U0B)h}|0H-L~#n_KCF)sDKh+b-U7&eZ;+=O(U)ujl+y
zIl9D$x+M@8i2)fF4CLBzjZ2T#5}Fzw3l_0vCXI_Y#K*G7kwY)V#<B&Y%_HYFl0p%{
z+oUnSd)TJ1oM_rO#y<5SGE@W~^cz8>UOm;z6^j&Ta=&@~S)c;Kmi@#_#a!i|8P4PN
zfg_0QrjT0`$f38q_oNR@*AKWcLTaYkP#>Z|n|T3zy|z=)nwbC%@b-OZXDb4h)%<8p
z*@)pq<%hw|EOr1ylkYSugSBy}by$VUKjCD>n*F?c7==}Vp7hBS`wKX<o}175#G7|c
zOz9%jP_wPvk=>f8XOHKi@x3t~trUXj{>~p{@jKj?!ds=iw|i=I5MR)6N@F!5n^4y1
zH)HF(Z^Z7b*D|S(B52590eLacXJ*5RfuA|g>gCVG);@;TPq#aaFUbQ^3;xDjiL=l4
z86;#xE(8VbA<U0Oyfv(j&!tRGn*s2=Rb(4a%Bj@Z1r-YahF~LomLNqB3~23!-xLV+
zt<w3)7b0u$8_<qPav|?(41iDb!RRE*;_!lf<|1$E%}js+V_;B00!j)8{cm*<zzhAq
zH7M!DkZ1xz#3Wx3v>N+|U{rt(O)xr~fV}{^zl@+iRR7igYG3XMD~bO<4WaQuG4+Cg
zuF23zLEg$hXMG<<K%h0eXl|edE<kF=lBDo$NHjUeNQPZhbcQQE9t6|54t@0#U+F+X
zlrgnzzOZpMN7N7_D^XdI-B9-bG=`uXa4ks|I|Yu6=Z3!xfYKs>WHA%Ux{+*fY_d?S
z+(sT=-m4}M#qG-nix2=1ODxo47#KDLWUOpUl~erQy?7njQGmoS*>O}o{0_ymhydD%
z^-P0G@B$&)!uwsB8dOj}97f<Y$v{HP(X~QR+h<cC_kec(QW;_T|BOory+CYjpdAmG
zh|t#w6%fn}$UKel6A=6|2UKIjXtKc+Bfp`MC{hs$Ac_Xoow!sXn^fVnRN-W-;xT}2
zV-*j>UM<4wrD(3B70p@{55#k)GUh&aBp(IZz^$|dB2huELc#OibW%VUf#??DYocpl
z$n~H0^LDc}QoZDvM*gqEbBOhDdsa`&@0#(pTl_W(G44>waoY4!bhVPe5^s$8MXDfG
zgD?WK0yCyD3&boisxbOtCi>pRVaEXGu1+ELhRIb<2&P99tR-)>fy;W+Mx^m6&;yj4
z_qbFcS?*MdRH1uD%#16d_xbBra!-zSEdoaN&JCdqb{N{6D@K4DnBnLKuzudEH8WH&
zK_>RF>X~$%PnCMP%9)B%ZZBl)3<71OU;_exRr^6)j6bA4A%tLuqkn|$);PTHc@eJ<
z_?ghVt^c%Y5whB4@9C8B?>0;0PzX|U@g!>Y*Lb2q;?nL9v>>`7p`A`q4a577jK4_7
zG*01aX>t*81b3l}y(QbflAGh4B~^TTS3kY?jTMgix;|<{&@tWvr9!8fFR<qgE;5bF
zao@V>{Z;&eJqKEf(A45nkXnW|hc9K>@O-C*RD$>=yxyj3QcMqVP)Asi3a~rE!BRl1
zBJC4Hhni}Pp$dB|X4hujhbj&#i4y?6CYN~p)BT5x8xT)kRIL!YN_czatO6Yb1Be=R
z0F~fv2|UVRtJg{vO}SOlP~Sz3`H$KQEg%WjTfm);tagsg?%4E)uJ(?90W8KAIiSQ#
z?7qExcJgAXiW(@N+rTEvrM6<D^CIdF+F7y~aHQRmi)^7?c;b4~V(8NP8}0C;O<;!5
zT7oq%j&27sRYCNLzts>7PUULYfoN@{CHHgvrJgOoNRZ}0(EWj_iH7WX1H+vH-iq)<
z2v{q*XGE`tssDRKuBBY~2pA2K1tgy#wvYHH9%aFW#ZKpARD+mrE69cn7+i@BTzq(U
z797-IEul>HyIN{UQB}>2KMCmNjozeKMom1PAX3FIHySE_t~ju6Y6^4slW&yFTjy+#
z1@bGT@PktuFXy6DGfirZ$0R(1MQ=+lhrpB;Zhe?-tE#M<DOO5(nhU=+7f3{{#{f*#
z32*1Q9mD!5bV0@+2RNjM2s+g4in#<?H?;HNkW{8>%5P&D*_}8AR6uMQ)hdKO^AU4J
z>#kXb*EnE?1dz?X_(Ef#Z5Dwz$z2Rmq@=$TdM6by7rfahy>RJ|O6U{O)x-35m0gm_
z2G)+~guwDm&7BU@EsyhytNG2ID=D$a?)%pSyA1nDVn(i}-z*`jQR{htz%f>Z#YO$y
z>9eCm*6?p0Q%&-|5hUY~cO9r?i9&hhU1}L-CxIUCYI95_^hu*sy3}Hjgrbnxk%TJ$
z0OkZ=E?j^aO*6+Anev|g<Np$UbN8VV?W06&OW|3f7W4O55`S4hQ`D4X{*6aCM-{wR
zZ;V6|as`Z0Ml2FoLH)~m3Mrb00|T08HIPD)g~vsQeDFV$)Fg#^^czs5zdU1e8jpu{
zFE2>^v?#slw))XoP7#a&v;dQqYm!R@h$*o6Tr^e|HFB<U+@~vDXi|gv#isz`1Bz0g
zXma^DZro^dcW85GXmTm3b8W-V{s$RU-=d1j;d~2H{4Emasw9A@C{zMRIv*@5caNx7
zWBQ7_<^pl>T7$9<!tg}27631!Ij;m`tC~krNW?Uguk=d<^nOkagd=r-!y?K~Ys_o1
z`nu5kh<wahVbG<_wPma~OcCk%8@vrI$Lzk&co>K<pIkS!apYzuYU|YdO|wZpvy9Jx
zbfAdfw0b<<)-q-qSLPTI8gUnwy)B@6&K1c$m-=eB&b6X<te_>8Q^dZ+)+!y>ee^dV
zc{7F2FmsR^9dp#$;35lxP~`RZU!*Bu4ExTpf|X5_<>4zn_#2QX<8|P+f08O3Zhx%T
z7Gi{gm!cUh$y@gMpx=bE2dzh71$%VcQt&^}<NKVk+eU+O-`ddUA<-C^NSbL+X~AGB
z06x(C0N?|FpK3hSz{}d;W#oG3J2|ii%U41iGZB0QcO!FoW?-&f`SiA|5Qoee4VM?P
z8nn(mzT{sn;#oto0TYfz&+)$K@<T3#GmT#rnw_u8p7{0Z70te#uzS~qo&6U@a4sz(
zH!X`RS)fNd!h<56^lgy-_naw_Cru#6P%L4`;q5e7Hbp*%zijxM<}Apq^Mt<7ie$X!
zI$0KvxnfEhmQuloCry$Z#ZX0X1jAhyQo=zuMfNmRsG<AYe})9KpLkbJ6=Qr3G)Uc;
zf1(WqQ#;Hl>2xIwIw@%H%2d<f$}g0vzaOLH8B}Rj9^_RVUU4Qb&79f_NieR;u1?06
zJL#C5B}y={V@|Fv#rRm=SMiB)+YAo67vu9zYrhCmA~lZl47{h;dM)5|b8675D`?;f
zRXS1U+^9wXrP+=BUu$PYXl%(coMa^M9}ccO7S1>{sn$*GtaXQ%k?ri9s3cFiN;9C~
z)Fx?QTAiXeGT5fTC$ziPx$)+`!t*~lzTra5=UdptYIRM%Rw^gIT>fG^NvTX|f{q#k
z6(QjyVM8MvrxBEl@&UcEGVGQMZoxK}bqe2=h`B4HzDkAFB6$xi3>|gE&H&!|?I&3|
zixQkW4`SgAx&8NlhepYhJO}L3Q=r?kh7p}#9xy+E!${yrAz(O&1(KlBG^mGhZs0Yq
zEb_tAM@wC$qrw42)ZHxRPuNNEZ_!0e3ttjA3ZSv$EV)U@Nj7b5!o?ZYAoSi;Bd#!_
zQ{<h7&joj}{l7}FD0lS|{{Ahz5B|z;#MK>l@yvXO^oK_-0*@OGZ+o_hRFd?}7eUed
z39<pd&w=GB0VeIMXizpRV#7joL{Xmx*^fP2OF$D!gqwwue8x4UhkK)O5p5ZZoDa{;
zvBKu7jPCn!(q2#kNdzR(p(704xk+!*e~34c2QjGmjvk$Gh2_iYfC7Jjs~`9MF_6N_
zhzNgz*F^|Mq9@Iyg3==(!<`DkUmu@lX53d$$lfhNvRpZ^#WKnk3nP{Xl1s?ro92ta
zSxY(3SL0$Ln@|Fo4g9>yHL$u*J8yW+2wM2W>Pp6p24O39iv6x>^(wshLuI&nx~afP
zJzf73!Qx4F|EuWdwtzxGhtwG47nYc+^i;t&Cpm1Y=w9WgoRZ4<0U&yz(NQ4`y;pR@
z!gnj#SV5(CMTrfAVl}kTN5MF<k!B-pp6+;J$5O99M5gm%*|KeOh$nc25S#n&7BJO$
zQ0a8@BRwrabDrfP9IdN~CU(j1MRyvMNVNl?*8T>K8>1p-G%XC?EqRRvc*K9l%)FEL
zXh85TEcTmK8FJncynpH?jPIov#|oN53sOFdGwKvy{HQ#`J`!nPKsmg-dhU?^KrrI(
z^u+_nFi*KeGF84XL!mky{d7O!-}~p;R!}$trbf51A`gC=qU_@En4CS?DXj=qtEwvb
zeJl}>Zy4!Z4;Qn4f0sAfCN>s3^EvS96AV=Do$<45^AE3gP%Q*a&WyK`+_vqs8XsC>
zzOQ4jLfG_oEO$1$yO|&#K2P@IgH|K&ZpM-CjkXy|Jq#^l7XH9Qe#!iNdHILkw_4;}
z<e9!%+lv&+gOCTHM8iGCdFmYVI)>GSzmit<8<Af_%;h7BU7CS=l`U1IyheXIzt#>%
zn`>hv9jd`bCS{rqC+5<-@u(8{-hom9_zdDuv>uDb#5_t^e<*?PYW!g~*SpYTs}-p>
z4e}GHbxkDWA#F{~$({YhcY5EMMm?)$m?Kgr>ZwL<5Ak$Iau<K8K1$9+&dOA7wc#oR
zhlLn?l1iAV)e@#aEUw?!=32ViamO;haN@|v!zF`*3K2iF+I@bogTh&(Hqts#`lyg*
zozUAc(4*1bjPi%WL=z?K5=;&nt<IQzwH{};#BcEiK!vEy^j!4b>V$Bmk$t~VB_A>5
zxFK*Sxpndu+cLaP-G=HIh%{JX3wt`cuG`khch#W?w&w<#8eTu=Gq}X(BlCKQ?Abv!
z?gB+?h<rw0V2Pjm!fM>=H;)4_bXl;R>1AR#S(+s#+!;(IW_I;|kb~}rKbvZFk{$=N
zud{rr>FMxuaAF>JL4blKOlh#_veuaN&(%oB?i~q}aFkL=;~XHPYll3)sU*Puo-4Md
z=LmS_pZbc^pbr*4DQXGMw?C#nm(*rwrBaOh8bC1Wt6w>OO&5Q?k^fwg(}HfGSt7aC
z*`a%IVpM4df?6(~`AT_<tD%U?!IR}-W8M%*m!FAB*gK@HkA2>e^~<u3dON?fZ^NKp
z@0V#WB61zm9kJ5=VLt6I$6=9}5({bu$T-7l`M<e|Sb#X3#@`-EEU%)bAjZ^uDC;c!
zZ`pRQqR2(`n;f5fq&ysZxYb(Af3kllnIo!i^N8N65PWB}Q54ObcIW;{^QOBE0dsI@
zvm??AZms41&tsPADYgFA5DGy_NqC(4_dj&L$DuO|URr{|#4KD&ABce6NfkC2v}*%9
zuVoK9D+fv_Djv^Y_9ty7M2JkmH-;4~!e)M8iKMPy#zD(gs^DS3SAF%$cIa~AeYeBc
zaFwgx%;ttEtecXc!1JNw%^;-fLe_mEJppe9<(Ye8(6N2-YE&X3ottDu{e~T`6{KWU
zr64L&5M`<74v<xTY;snthVhdX*q=6cN?NM$TB?vZPSNjIT(F+MMrmc5`O3)!&g5Sz
z80#W^6sv7saTc@WPTs{inx7i-OzVqhpBhkh!0>~6d8I4E-!|i|pAY%Vv#Y;c;EQ>u
zULDg!zFMV}ocPleAIY13kdK}o-kVrBU6!-Iy~ra;#uNN@W(>Lh-j2SelRGuuF5C`)
z2rs?l3N-6?W2<*fn9gh=Zi_u5jome065I}6+GTcap#G-Sgv{E#2ou>9i9IG7?0yT?
zz8&w=HM`vQOS<gHvOrk`kp25dJ`J!+JP@H;S)nK1EXYMfva4$o|3Qy5PA+jNme!mG
z`KHI1*R)9ls{!E=4-g@R9oy=$jiJ<kz|j}Mht)DP*L~NVz5E3Akvg|g#JoZFh)!;E
zA{o!Np~{l8>MtxbCoPzZ+%AnO;ZD8-JhBDrM2zxWsa`{V2i){BuiD|uM;O!WL*`TJ
zk8<aqZc@8zH&W7NE8S!EKf|4qsHaE%!1wtf##l;<4CJr8F@5d@s36y<)}ip$i5L+4
z@?m?83JfgOqbU$C#Xg?xy~T2R(jhONw_RoYxzq^qmV%YbiWAt;W|!xNlJ(Lx$^5<A
zdACj(Y3Snvy$&(gj-B6pw(rsWI<amE!rEgxj6C7VdcO9y4(vsAZVs*~pX#oCh4rQ9
zVojzpg?d1H?)fWU%+!q9hZ=P1Tn!jvSg$D!#6b**EnFK_N$eM8DbrEi>YI5C7`r?b
zSt6&Y=YacCa%fKe;@|{39hiKtVT|ZG!T;6*tj$_6ZY=(2!g(;;>*>`yRn+zlFoK!K
zxg&~~Lg$WHK1_0q+1|X{e3WjviOEeSG8I$hG^LnrVUOgORm0=9MT+^%A47ZsZOHeN
zlyCI2EE3dzp{Vj3w(p>a*{q&P$LVO|Wgy;cGd56U8tbxmeT%u1{~L4b6VH`k+&xAA
zVBMUpK`ZZbfP=y|r1y<Hh01pr&dBQFUV0?*(M{mSFOHqRx5h(^Wy^EP8O<VFclYja
zm7jA$>Aq!|*FIV9qE+V`kN{I|hB0L6GN85mGy4+L%mx#Bk}U>%apU&`1wP#QePt-(
z|4y~F1b#fjD%Su%6HjIWA-%pE5Yp2;sbt@dp=V5O3WZR0G)B6XyV9VI|6MdA&gq*T
zB`iN+Npv<AU;=qWw)Y+kBf|a*btVw}cxQBDgn49dY=P%47U5Qd{duHk%Z2aU4RybM
zA+^?o&asj8HR%G=nD?1*RQ{fZDX=@q#0<_Dstv{y$kLfTX(REXXl2A|AZqCy!MeXQ
zhns=5g1Qy4v0B{PHV6pAEU?!UL=K#OAs*E7Sc0bI3zYXlEIoU!69mJtV}T;=4JsJd
zr$I1vjGsfneb`m%zmIMbWFo0%Mg_sc<=>GxFO&9<=)%p1WSQUP;wH?<XdE0D6gZi_
zfa%a4qq~_a`mZ#_r5tmp%pk^?^%Q5kp?7%o921YCQiIK7Lg;j`^yKN9VB%luWv`th
z+vQKNeFHwUI;evYdbb5mJ(MxG?phtLtp1aq#yDV^7WH0*Ef>DnR|fxJZs7a%Ob~Cy
zXqY!1#lLYG*KB*amZ-!3r=FGf@)|Gw>polT#nx`%Or!dtFC)WuDd~EL5B9cB^oIMK
zcTZ^4M`)4n<86|`a6Z387r@lfoG7<fDFGEWO?;A4_N~MQN{$o`jjWGo>XB@}6VOFA
zB7QrBXEc_cdxYCt>%2{swy6ZkOw!ErAK)#j&9Q89dsDEE*JegDHou=dQ=FcO+?X7(
z%OgBu9sb(7so2G<d_{N+RlQ)tdc+z^o?x@zJmj|?h*6@?p9oeJ!<I~r7OfusoGIyj
zef#rzs6}kPcRDMN^f`cpQA`LbY85(A)))0;MWoI^KKrZNdq<sswbINnHe~#7GWg6L
zdNniFIvQTHQ7%oUI<MB!?Xuf=MWuF?dY7ID+R3Zonj22MI=I@OO{K)V+%IN8PhBNx
z4dL?axj-`W(Ko20ynn2e5vtD6YcPBMG_ZsgfyX>Qyj`a(bB-Mg(GI1%v07VQb99$K
zzPzq;aT@)AueFr4%K9G*EbHQr?JA;+kqI*f4;vM&lS54sYDmfEJq<s6>KbD-o?tUr
zsh@eJ3%l5=&-6Qzw3tnJU^TYylNvI#K=Fw6;;y-)1I!x7ylzKqukTK!pc_iJiGcd%
zQFcWk=3TO3DQa({$nETJ+WONLGS7V$7L)p<L{;orPS4_we~;tx3BALFo3ta5p0&Q%
z`H;<rSuc>g99-=`o^<IM_v)EnGLh3(sU<BtWE95%89z!EK<=i?aVci61MI5ew=YWx
zYfV%zA%bHzaO`Gq03>lNDy}HTm}?FkcE${?g?%Yb4WQ+eWh0JOZ}xC->Pl-rBs78C
zo9{bnP8zGNr{whyKJ^#`^7f*hJ1+aT5Ex*7S3g{3TRKqxTnwgFoX`QGg5$6g4jZ+S
zKX6;?FK2%|L_KL&e+|Ty*X*_OwSjZw%d^(eT|Fq{Lb(<+W7-L0Qsd?E|DygFab{lR
z7^b&Dc;s|vIc)iI_M&J>SVNKpZ%bw)De@W4hN1GY6|0qe4O7$*^6LG)*hxW4n5G9_
z+-YY^$3ovzTi&!m&1oxdp-l*^|I(Z!-*r3s(2W+)HAP{HFIX<asrFK;pNJduu=cb7
z8rtpo4)Xo&RNmxigjb?m&Gtmo==V(N!~7(%q}(C*TNGP+^_&jKK~oCIq`Ng$8}9xG
zuWi4k5?*Uy0=T7k_kc~`8gB{}m~kaR2-+!M>gdTgzonx&irFVJ^Xb$jls<Z%CfL?T
z#39cWD$g~rOE@`bOQZ68d2Nle>_^yvW-VcMT~GOmqQk#hj|W<%#9X5)D1E|r7I3ng
zct}R7<>HtS(LYmk<SwWqBuO50Y2XZlk~y7$hviT4y$6387(7g|_#bEOQ4@Qt%;%Ik
z5JN-MTU%nO_;f$T(Z{g?dA;E;aQlpKL=mUo+nb#J;#D2qw-6jZc3Pwhw=x~xObI0}
zaz4xzf84gC=UEoV{7!z>l9npDYj%)YLTCEde?}>bn8S-Uw2naUCGXk2GH0$8Wt<gj
zaVmv7KNLQtGv#4^#L6f@y@0m{H($ESRXhn&jWFmx-0PU|gGK+?@*uS|nE6D{j=ejJ
zbm>4_TU~L0@r5Vl@|9)MidoPbh~Eu{M{I#S+~r2K1Il`kObnmp0{*h=RGh=hsdp}z
z8pRCSh*=B-)mUwcRRvZz8X|mX0wd6evHMxQBJnpPz`d6LV#^rXtV6}(DCybIiy48h
zA8XkR0l>A5i`1yNL22bFq`d7ybB`Uyifn)W>b9Tue!I`*{i|NmYDKr^jotbsi;LR`
zWtlb432zhceYcsu?D{n6rmJU6%O^J8`v9l?*(FPxj>}cM#l3)LCa{voG3(Ci%B^<Y
zfgCP_3jbkZLkfOve%bu7(-Lo1L&=vX_-vTpb@Q=Op-KM2je(uDpYQvqN%zJ};a{6~
z)B>+ev9D|X`Y`);McAWpgDsaf4`2P@t>CnGDpq<vVmeBNrlIkvy;I1(uec*|YHlh2
z*_tSD_f9ZMf0vKL57?sm>QHl=T%RWMkN2H9C*QQSbk(2wvX$qKZ`YUD{15VhZp+}y
zJ5yEofEW*EZ6%|<MQiif=bil>o&oeE4kMS)@M)_PjJh`W`XR>F`8-APZ1<f*wpy^*
z%B|_ri@E9VS!FyQyn)-;S%vCrZ^$Y~^oJVtE7(-xg54`!KCPbL@-mqwhdvp%1(gLD
z!GMGs&SSe=?xM4^TII%M@+O&1_NkLU)!()~A@`m<<*b*VN}K2V49W$nX`U8hz(r)U
zX`NpX<U!`&oT#r3Bf;!gUQ5v3w!|yED{oR#*Hu7cb<=QrQBWXbCSol7KAXX2Qd}Ul
zIzpp=gP6jhy$fWRpqL$|*l9>fIJsg*_usePCvC2ETA%wmhO*LKRjXmEYFdGJ_rANv
zw%4?F_pYLQH6Kg<k;Sqns;07~pdau;*of&F@|qjdblY0WNtXsbnM|vhPv4N0X0nnb
zr;lw<$k2V+2_hlqg5I+wSbCt@s)1h1l9jj9o_T-!?D782-gK{Z%~6MU6d9imHmrwj
zt42sY_@L4%k*f_D!RC;09?MptDKwP@n3MmAh7zWU)tv_L4MX{D$cGO^?#u6;!Y#CP
z=MQRC+9V6!(a}q1u~RZ$IplX|Jr*=9_YP<~l~>mO>>mrFVnFkF2Js2m{p@sirM}I)
zzk}}T5vOF>t?#sYHtTN9{s(PFhmhp*ZSw1+qHTGIX|a>&(#S_NW<^ledQ8^wlW8aO
ziCY=9{Z=VeXBnlC4j3w}-_x&{u!`)k`{!orxdB`UbGx-G%JG-W(n9~3e%Erm<0l|w
zyiu7_P|+SOm`{O8J!qBxix;ZRF#L#Ir)a6~z=gIGomr->5}u-Sdx?V;j*j`%;I{D?
zuNTyc<w`RxaiZTtt6$3kafMs;z9{;^o<>}RojiShyJMA`x6%#Byg84pu(p6=YRV=3
z)Yi1@(W}B8aT<#uSK#mSk^KoyRq(hSyIpx@yxXc~=i#bVxrZljp9E>J(U<dzOZFfq
zq1z7Y!QLrOjWvbjU0WS}h50G(aaa$p-zIdekIUUUk1I(fEyylX`aA3}c7?Q4;^tIJ
zOq*=83gtsGk7cs>NSD1Qi|xej8j(F8tU1HyRVbGAI`?c;r>DhFYeC!`@c}41tK)LF
zYZ7#<{68;G2@VL~UGvr+6`Rl14vg_@H4lT^R&t+LUO9~C@zHl}_XAh=`x#X)ft2L^
zMlT`n^)_CcOKoy=*hgQJV3UYHMEPG^U%fbUF8dUdSC3qhr3rm@11mZs(5Dt`Py!rO
zb@B1)j^n5BwS@CV=Z>lz{neycl^dU_Qmc%!Cvz(*LJF~sppok|P9=lpTNQ8<Yrxd@
zPYcxAs2X6;g&#y)d@r*cqvQJGATsqMiZKE9QyWe(tpv+Gm4ioKUp6Dv+51Hu_lyDX
z>i&f1eY<>9NiW~s26zxH1y6hE)AS#DKYU~nQ0M0fD|5RSo2SMb!6JJGRITRNOJ-~?
ztU$hs_12q3knc9*c$jZk4mfYR$aE)%^JDW@wlnVXoe5(gcW3cP+Y<4}D5JlEN&&Y{
z<I^B6S$>kOGCJp~GI$n1qSjTmdy5u_5#)-P_Tq|S$KlVBQh1-H(*Ffgq(8p%jB@H1
z_byY1rfM!RJ^zQ8*JKL+*S+5~PipRCr_SZOmc89dr?&UG<CdA7buArEA+aK)Lg<t}
zQj3{d6dK^2!|eL(6Gr7jg37p1AQszj4~l%G5J>t)Kngm*Lvo30caAeUo&kl$z{F)F
zSn~8CE6u7C`SsX?Aq@v!c6ak_c`U{rYoE4{LgJLbgC9)Fcg?Y2R^QKCc9p40vWr*|
z9a@I2*J0f0lJET#8+m`SQ+^<<i{JTd`wVyy4+|n-5AfF<`hnXy8cUnq4(nZBU3F0e
zG5bkApzY6_p#_)y?tasWz^1f^jYfUo5!ROyBaUnI<7TaKlDBE-tO?*|?`EG9q#fe*
zSslK}gpZJlbGcCL!jIVaTRPV&LEXkG>Q96FUbU3;Z4yf}lc9)LYoC9_uea_ds7zNF
z<u7X6yxz5s4lc5NylZ4gsJ%Kwd{m%q(GqU8fTP4OIWdl_2(U#lJyvx^i>H(q`%t+|
z(fPiYe8;E>+sfv45N$i+;>j`6qxJ)cUQET04x%N{x-R9iSXouFaOwj07=Em9I^lWh
zRlrV$)*)85h>_|1=Z_HjaOb4Y0;q(!=N!3~`Rv@v8StTu&`I0e5lsg#65%71s$959
zmoWTxPdMWq!_GQqYq|8N`s-IKUwf+$@yw`m5vZAr0t%GeM)(Z|{D&#*bXktrv1SoX
z$U*Uv%8TY(wT;J)U8Clu6;oW=g3GLO`CfjGsgvDHP;6xKf+WiOyb}$+MTZbXI!}%2
zZ+n-ZIwRP-xQ&iQLSs+M)>9P+i_HeMYX<&5A+;#o`xmd;K(61nMVAbYMkFI_vMBZ%
zNXr)%*(q|m#I1(1bUBdSPCff*q>82$ZHE8#vTM{AZGQ}$7SppcQYf#Zj+kz!oh2Vj
zLy$R)Rw$S-^+Qi=ba9pSLN~aw`})UDJ@$aa`S&m&fuZLeucNwCx?e}l!?8cu^16x)
zO!HQ@`Q^ueJ%immj$O^3+kZ8WGF`bn5^Yusube*%4SS9|%&IL59Fb^$o?8Zimc-UH
z70xV1)mzA|h3g3qJQ^c~_F%$;^vFJ1$Wq{mfXp{sY+wzcF4>Qe7h?W>QA*yy8#+Ba
z$=BgA$vbeS-Q|)isKfFrM~Dp=7egNz#O`N~N^|x(!7Ku&D}ir*Or3cVbIc?*&x)R5
zz)Y4-Q-8m7fbb3I<yH=mv;+#VJy`Iob+7Vi_VZ2({BzCsGMx^Sel+Inz_U4K#|pyx
zhaH?8i09L7S7qJi-dvyP&@@1gXd3XvOAX0KRlmYvDcWz^6Q#BFRHp#rI@6jPq9uId
z^9$GrVehH>Vxo2>b)oQGd(>{V#PHXjXGvWB0FYoJ+<^WmLDL*qOhoraRm01DwO(%L
zrSY*~7B#t8BdyTmL+u?JRHaf#6-z|>E%(lZV!_aI@neDPY!p>C>eOIQQMJiVQrV34
z`sJ$I@Um5H7l-goPK)(~l;`D0R=XotO)hA9s;QvVPUD8?oalAUu5?yQ3fk&WR2vE(
z&;0#&{j=Q6M-!0Ex5s(hpTkgxKYJFw5fl~H(Y?Pns!hkT*c~Ni*!10PGd7eg)OBB_
z-7?~8?lkjAvvP4rRpoyafE@rl=3V%M*S{W#U2E%GybBxg>4y5PZuO^)wkq{zL;fx?
z`n9S2yvliFc4gglj;2E0)CQk~{IXqn%=iiiT?)f#2ie*gACQmi%k!~3kS;}e`1x|%
zbd9puirYWq5!X77e>cJAzwU6mHWh-U+*O}hKHN0*i;ADRtkC#jVL6`SD;i^pIb8a9
zWu~ibV(sw_9ukn4W}6)?A~Za@r0GcZf^Ea|Ie)d7l!=dbG_}|-{Q_W>PGC-Fm`Zcy
zy-ikk<d_K@2FyN7j2d9aMJ<Bv{mRm~#K<5P%iPA1<sH3VL`$~GgR~30!TT&;US=CE
zFXy(t3w-L^eb`e|5d<gAB*s=;)6~Ijii4fw38~xI&-yzSrl8_;yt*|Hh}Q$1EvAWw
zpTwIL45=kDaZa!zK2966u@a>Zh*5RzMpl94;IBf%w%P>H4er_A?2{V73yj$MhkSUC
z$w+`z#)B_Xi^O@SBDOqIezy@xTimH7+*Ge49w_1#@Xc=YPX|i3eFjHOzbuIf-OW4A
z(*wmWLvf3)8m3q)lC|A<M`?9XeyS)LE^4fH<kIsG<SwFCPU94xxuu(m0LcP;omU^m
z6aGoY%kTA~4#}fO#T^B9@J=Y|4odHPT2ARJ2xk+t4X6J;OjSpn1KD2LS2@U2=PmJt
z=jt$<UNIYtyL6+58`=&twgM&n;)$x>*H=f`W6ipv!lcUp<{0V|y6ZI8e<#V<X$$g@
zNzs%Yu0P3UDPbiTBHY3H7nRXb#5({uQfuD!UbQ}F#?MBeYV_>EjBs1xVKe-CV>ZA}
zXE0yu7yBcK|Fw~jmsm?w_8THheHpKej>^Ot^x`Ssfk(QU;;H@pbaV9#^s@DeqdKzB
zyjQE4c+#qk_n=PlF^8_FY)dGnxXk*A)3I?6xv`}|e|-BS8Bdozil}e_UMwzc%FTTV
z(@)3W=CuVVniRBS-rtJHTczs=4S=dTa?431pC4tJ&(Je-XJL#5MGhA?e`TM_ZQ3Ai
ziA_K1tZ>sEW`_-O(ABOF$`^B4m9vkB(r>Jfi|xpDX!cng7_?rp-1752_ULqXm4#i}
z@^nXY=L0Z2LbsVUbn@k1yYKETlx06E1nn>0u9>QZITRl28PZqH9A7_@u)QQj=W0W1
zQ9F9oiDKb8i{fC;x!jG$wr3ZK56=_K$5LWl9Vg27s`#+P>Ya%2*m&S)aGZkOBUIX2
zrkxC0OYiI6Z3R9(-;o)WEFafo$SUcv&5mV4jWY5KrdGVwiU3<y&d+YA9j=-%^<m#e
z85tJ^JuOwLDT6+f2ec>f5nkNiL>kYBxb^*#Rid3a2`dTJ0A$eZ;n|Ov`T{2a{^HDW
zOM^h|#Sf+{uxq@t{$62NI2^J8`$LQ0R;LRpQ-QTaouG#~%5kp#7QL?R_!PnI`7zwR
zVadw}*Q=>Ix9PwxwA*w;^6Pq1bBRBPJ9C7@{xWGg3^BH_p~bhV*<%1d<Ic<_pVFzI
zU4HV_Q?mYO@l8*iyr2a9apDbLH;(+CSt#q69<s>BMi}zn44B-mq#xtQO0YdpQo_a9
z(HKa(59h8Ps`{ozkC)sW4wUFKosmVsHy74;HuhV5Z3qOa28MUrUXyfbs2CT2CVihH
z4Ex2AebwaI61q-Ir{57Dqn^YT*0vj#nOd~gwS-|lrfrisR?zdh;4@f#%C*z^Gvsg{
zK~~=#8@18>7US6|_Ok3qW746d)D)@Mam-yCRorOSZMD|`x#u<e7Hch;C1%nQvguEE
zY}mx^iMU)IsfvdeVS1{McW1jqR%$(!TsrOil#h#C-#3n9)SCWy`D5tr7rq<E@vd=1
zk9ER#hKn{TotsAMy8xF;_)$LSPJD|q)T&klcMdlm59QChc<VhrBMXAzMOnTff%Ut5
z>_pz0@~5IRncUnPnuF&f-n_G=TxwE?li#2Z1~Ul_^XgX;C+&CptnMe_VNHYsRNckH
z^Vrg4bZrJ!`7AkC21euCXX_RsDM}BJDpTW5A_&>38H@8o{cjHz$Yz&kXE0Woo6L$%
zB%!1V>DH*$U@~yjKuZNRsMHBkG=yOM*(18m4FJ!dRlN}6xb+q)&vJ2HSm|>7GyRkD
zshZf8n!C=jXN1{}QR?eS--B=4c39{GH&Dt-eaqk<A@1*ytcV1up%*nj!0VuCcZFPw
zLFG<8w#NxLU*}Qdjun0~z5<Ufq;B)NOjpF)1frgd@Z_DCt$#qqfN~ym!F99OEbU7c
zuA;+c$j^U|ipq+*P1Dshc%D{S)OD3_JzgVA)FfSuH_)9{kIE|5&}B=wX3G_BhquX^
zHt9?}ZEZm=!=~*c1e__Em=YR8`C4|N@c!?@KlY5+jb$FW?)?b^<A!1_mwD6&;+bfN
zCR%6koaFoc$uwZwwQ9QWK3$i{YMv2)ZN={KqHX_1mH1$iG(DV;Z(Oo;^39shca3fX
z!83>ds{6tEkXlpFV8r3ZENFXhhhJB0d|`PBE5~g=FrUHjI)YgWRmEm^t{ZyO<!)PY
z`Cjjjv0ikW8q|ufUnrgMKTjt|V%Hk04)XM;;CwQ#<BU=nEk$$Acve-S)g(>1<NlEe
zDTG@(D2Td>dlY}G?$-9+{qA({Sw|;?j*)N9y0mPw|0mL=;|EiXZNu|{VXBKorcmp~
zztpVaj%NJ}+S{JH&Q@Rzj={X8z47V2*m1tsWk<MBJNH}mB^x8Ffg&@PqpCQ*UG_Zk
zDUXw97k;r3t;k73$IpVdzD8P}TG(QEr@w2$!xKj>(o#fgx*sN}Zq21fZ|Lt#`W?&e
z(vE$!7o*Y>x_xC=z8MbLKyIHcxN6<-Yl#b(7syzR3CI7Oh8t4)%@ffAiB@sQ<a9Yf
zJ|V(2onC+ZUJulgwJqAij#Kafo}-BwX?aoJ=N#(7STmFH+avkr+xL#~D&wbLo%r|1
z>r#h?p?4h4tz~9sv|NNZ@*pe2*nHu!K{W%LPP<o^n7(vDo%<6@j;@TBxDzpQkjunc
zR}QLcGwDUFhGrHkWMuYZ{o@f^RI62A!$m3FiG0Fkq-2(d+L(sJ<?gHfsL^0r#OQ4W
zlVcr!{iP?!mLbpGGMC4_vm_HR-`>bYX_vhXa%h}62ui~|s<e&f-Rxz*Ej3_;y4o!*
z>M#6D-FK?uFV@m5@DW;;?BsNb>P?{RW={1VJHGJFXJa~VtNN)|UPB>s92@^O6)6|5
zhbrZFLeGtJIO<pic}kTzKAnWbIP34n2I`Qh{<XtfZmiBr=?~mo?7|i&9IXgg-LN;c
zts;2Ow{IVT<IGGImiXbS?qHHBJv3r}sF(%M3oUmo(=IYc^tIKVKhyhN=+buV;t283
z)^A!m)o|B`WPO)C%v+ObFoM0lb$<NQ;bD!A&BFFBJ8K4?-EoDCc1sHX-`Vupw$;8I
z!|>5-RbkhQ!5G)Cz%j&F#cEi-^gGdZR_D+%jyI0ar>FNaJ`%4z{NW#C=TF$-Qm?Px
z*~Uj0&liX0AaL*&O+h=`N+`xp>+ks%4j;K&%9B>K;5x!ngbCDOa@A_$zpaUzDwgJx
z>$`1nyKUSSXq+FKTXB?Vkwi`(JzPJYjr6%}8F1m;7kB)%6A@2)F_ZNpf&M#Q_RVgF
zBLbmyFH!EJj?H`ERBPr>nclo6MnfkT6k@ui9>$1MsZ-SZEsLdIV2hiTclQv%X}7C8
ze3X((`upybmZ$q{U};`!_0ZtW;jzQ)yvI}CMyr<szgcQBhmot-V6T%^*I`}S78}Cm
z#b(n)i-n!1r>EC~Iaj-^72K#WjE!wsEwA%EdS8fFdRm?(IVTGBl!mUs5&!#)nwNYU
zInuNG`{umr@r3K9K`sS#ak+TEQuyJ$sPx{tV#kZ|jbDMfU%knB>N?jI0E4RDk{t%b
z-^KuBi8z=@UVSRz5O**g-k^VU*(t6en1rx-HCRIp+g8p2F|COcAIr3Z+(jiJN(&-T
zg{Ck@7DSMaZ7}lCQf->{riL&&;Kdgy{k2I;vEliD^un&Tv+5prW@1E+396R^S15m3
zGRTdI$Nwa|Vgob*R8;w_XXvNKz{p^@Bj5C&<fzv=yKBwpZslep)JYmwOhVjPqYHab
zjaiTP$cY_+6r0S%^>VWI&|-Nc7{Ot;fB`aJ+ZGJ~2LRBF{)D8!%k<_LQVczqqSd9%
zgz7gnO=_xW_pb733l<yA6Rfs)KQ)`X)>;YbIAW#0s&$6v6UYJRa*4<C-46#jqJ}GK
zSBJiQ!?Y$|WneE$b<k;|)R18pl{1WuXVCefnC<Bzhv?`_@lL+>CtlnAKh<XOT`5)U
z%wCh8Y0a)Akd*DOKi6v!c;YR;^^L_eUd~T^cyJury=q;dXatI)Sv)(~PxzW5x7SL5
zZ2D#Al*7|Q*vd!Zf*g}YK}+s!YhdR=01z0EOb<(J4mseaYWr4<_GTt5t{D&jD=@CD
z$-x`?v}8WmKlYbl7|g;kR+E9NnmQIf%`@Qg^FZF7-ehDmf!JbR+R(*@#&G)NR8LnF
z1sm;X3EgV2RrtxKQ<44n+rKK|NOAD+!?#aMo~~BOyN#>Ybh##zSrxO<Gs)_QHxm^<
zh~CuTs-_8-#|n8C<z$4Sw{bPy1^?!JVF+c(!%<i6aFK7VsoZ(cGXzc9ru$qQ_nR@x
zTk)E6Hxp3MJm1(Bjp_EWyfP>3FC8L~F!ALhyqqt8M5iLsX)hZ+0yGSfv*ZVnF63Xo
z-v$&EeDm9L0p|L(jjuEHi+L&OX)_IV*a0+4K_z{@)-9gJ#>;<ARUMe=*L1#9BCF+P
zizdh0Opkw~d<Jk)i=_Y`5Yo_KE7#%eMbK@6t--F!`@N-*>n7TIXFX%UR*RP1++Q6;
zjWTI#zGWZxr}A)@k&`DG_!AaK_C$3Yx3tCSJ$u{05M=sI_6l@^O-njHw7H0;iaSck
z^#IB4zh<B?6pnORa-ruJ!~r<Gz_*6Kv@9+Tj|mj&!;!wQnue2n$Poh$5ux*~Pl5h(
zzU2S;c!B?(QF=Q<@!uy8<or{NlVu{Vd#kCBzS;z3b>jp@T}VYU>_bUPGykv4=V0Jy
z0~~4B{~T&!76iSdnZW)M`9B{-@H6Jd;_bl(Le|WgZ=)cA6ZC(p+3x}hq!pyJOJX&O
zV>QZRU8%=!)G}l7a$^@Du^K;-!2ge$9x+!>WhoTqRF<LOHaQwLI_Ukq%-Ce0L;z@A
z82ccMl%JcVXy&76mh>mS4S)UZEEP)A|5+}bu&;rJzYdO5#Ab8Qc#mHht0*c}II6@;
zm=dQr7^euFt_=+<k9BK%j}IJtCZ@@q!C4(LeJ39}!+rc>&FESMy$8DmGw7X-X^{|+
zY7Fh?{VA{93mOe~5e<42XE6arcU|Y!1@v_lWQAzX`clVMyU1?C-3a9ucC1pbn#sd0
zI+^7i-{Xk-b+!B)b`632iEYiP=jK~GxaS6?9ZvtzWV2-u))oI$Uj=5RV!hq1eYeN6
zC{Ledtb>Yfdxfzpf~nmq>{u&<WZcwr0&&bA?SXX-fAXcPz3X=^tv96Q49mAh09(Ha
zxqN1h=-tvDK3Y8njw#Y@NJ0jy2ZC1Afhn{Gl(e^!Pm4h9>0^LfP0M81u{n9#g~#sD
zdGgrHNjocWOl&t$n<fEuvo+-$+ZSMl_e?uwSOj-}biBO-iUVwI(oJZ1x$^&<L0PT(
z;F!Tm@k0Y8q4RgW^aM>3R#$4}$2k<44s9g9f|}C0^5PDoO=p_1v@MB2<aR&r!8t2=
zvpiS<*wrt&ShTq(teuMI&j+wa&|>%uO&oEG_n$fG_rJiY2v;h#q?-e@Wi{dy>3t%7
z_{|`LzM?fC>E!=wTQb(pje0FyFnxtTW8bE(v_Jl*wrDJ^V|;ThO5cn<+Qs)($qWnU
zY+H^UEIYzAOT#AYI5}rp&YgD;AbVm8b~iCyRLS%VSJYsun(=KnvnuB%)tEBPrMrL|
z&da4AHeQ+{k{Rmt^nqAoVbp8?o(EHWy4lx$bxdH(yPwk-bjsLs@y&H?vwI)sHvCnw
z(@kdx*H~)C{Qhwn|6YezOPLwgmUe!uH#ZR6?y&sP?@%6*?6uFOf6G4<Yin7i8rzzn
zc;WsF#fja!KV}Ct<!(Bya^Ny^f<#{ZZgt?=8u5d`wKXf8>X>6zOv%drVglTioozHF
zh^vA-V(r#{{VxNwmMvWzwCcc(Jia;nYuJ`CJNIVn%lx>cP~*egGV5G*Mstb1g;mUd
z8(DoPb6tqCU$gbyx~=b8fB*dtTI;SkjgMhz$iAReQ`4otS8r?*so4))G1i#twUW{8
zLgfK1p}EtR<$cLK<Tlg0@O#`%hCikkzp-p6wAsMC;_%uJ<qy7{_G7*#`Tw?Lq38#D
zW!uw@lYOrLG~ICT$8KB3mlne1Tfev`to&G>*|qGc#ex$ar@0;&D(~Tc$5UYYL2*KQ
zi@EW)>(a-6&sg_0T)8*9(Oc-pyt&TF>laS(ZcqGd^f-U%^J>4N-%oFN$dWMeV_$98
z?(6<emz9EjfaOJs#miT~t&@%dIZ7yJqF^huP<EZ4`_G(h+p<?_o5~*Gp)U-cu6{1-
HoD!M<#O_(#

literal 0
HcmV?d00001

diff --git a/01.fit_a_line/index.en.html b/01.fit_a_line/index.en.html
index 0992be0..7dd703d 100644
--- a/01.fit_a_line/index.en.html
+++ b/01.fit_a_line/index.en.html
@@ -205,19 +205,41 @@ feeding={'x': 0, 'y': 1}
 Moreover, an event handler is provided to print the training progress:
 
 ```python
-# event_handler to print training and testing info
+import matplotlib.pyplot as plt
+from IPython import display
+import cPickle
+
+step=0
+
+train_costs=[],[]
+test_costs=[],[]
+
 def event_handler(event):
+    global step
+    global train_costs
+    global test_costs
     if isinstance(event, paddle.event.EndIteration):
-        if event.batch_id % 100 == 0:
-            print "Pass %d, Batch %d, Cost %f" % (
-                event.pass_id, event.batch_id, event.cost)
-
-    if isinstance(event, paddle.event.EndPass):
-        result = trainer.test(
-            reader=paddle.batch(
-                uci_housing.test(), batch_size=2),
-            feeding=feeding)
-        print "Test %d, Cost %f" % (event.pass_id, result.cost)
+        need_plot = False
+        if step % 10 == 0:  # every 10 batches, record a train cost
+            train_costs[0].append(step)
+            train_costs[1].append(event.cost)
+
+        if step % 1000 == 0: # every 1000 batches, record a test cost
+            result = trainer.test(
+                reader=paddle.batch(
+                    uci_housing.test(), batch_size=2),
+                feeding=feeding)
+            test_costs[0].append(step)
+            test_costs[1].append(result.cost)
+
+        if step % 100 == 0: # every 100 batches, update cost plot
+            plt.plot(*train_costs)
+            plt.plot(*test_costs)
+            plt.legend(['Train Cost', 'Test Cost'], loc='upper left')
+            display.clear_output(wait=True)
+            display.display(plt.gcf())
+            plt.gcf().clear()
+        step += 1
 ```
 
 ### Start Training
@@ -233,6 +255,8 @@ trainer.train(
     num_passes=30)
 ```
 
+![png](./image/train-and-test.png)
+
 ## Summary
 This chapter introduces *Linear Regression* and how to train and test this model with PaddlePaddle, using the UCI Housing Data Set. Because a large number of more complex models and techniques are derived from linear regression, it is important to understand its underlying theory and limitation.
 
diff --git a/01.fit_a_line/index.html b/01.fit_a_line/index.html
index 6014986..a45eee7 100644
--- a/01.fit_a_line/index.html
+++ b/01.fit_a_line/index.html
@@ -201,18 +201,41 @@ feeding={'x': 0, 'y': 1}
 
 ```python
 # event_handler to print training and testing info
+import matplotlib.pyplot as plt
+from IPython import display
+import cPickle
+
+step=0
+
+train_costs=[],[]
+test_costs=[],[]
+
 def event_handler(event):
+    global step
+    global train_costs
+    global test_costs
     if isinstance(event, paddle.event.EndIteration):
-        if event.batch_id % 100 == 0:
-            print "Pass %d, Batch %d, Cost %f" % (
-                event.pass_id, event.batch_id, event.cost)
-
-    if isinstance(event, paddle.event.EndPass):
-        result = trainer.test(
-            reader=paddle.batch(
-                uci_housing.test(), batch_size=2),
-            feeding=feeding)
-        print "Test %d, Cost %f" % (event.pass_id, result.cost)
+        need_plot = False
+        if step % 10 == 0:  # every 10 batches, record a train cost
+            train_costs[0].append(step)
+            train_costs[1].append(event.cost)
+
+        if step % 1000 == 0: # every 1000 batches, record a test cost
+            result = trainer.test(
+                reader=paddle.batch(
+                    uci_housing.test(), batch_size=2),
+                feeding=feeding)
+            test_costs[0].append(step)
+            test_costs[1].append(result.cost)
+
+        if step % 100 == 0: # every 100 batches, update cost plot
+            plt.plot(*train_costs)
+            plt.plot(*test_costs)
+            plt.legend(['Train Cost', 'Test Cost'], loc='upper left')
+            display.clear_output(wait=True)
+            display.display(plt.gcf())
+            plt.gcf().clear()
+        step += 1
 ```
 
 ### 开始训练
@@ -228,6 +251,8 @@ trainer.train(
     num_passes=30)
 ```
 
+![png](./image/train-and-test.png)
+
 ## 总结
 在这章里，我们借助波士顿房价这一数据集，介绍了线性回归模型的基本概念，以及如何使用PaddlePaddle实现训练和测试的过程。很多的模型和技巧都是从简单的线性回归模型演化而来，因此弄清楚线性模型的原理和局限非常重要。
 
-- 
GitLab