notebooks-3/1-hybridize.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# 命令式和符号式混合编程"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-29T16:56:39.722075Z",
     "start_time": "2019-08-29T16:56:39.434870Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import d2l\n",
    "from mxnet import np, npx, sym\n",
    "from mxnet.gluon import nn\n",
    "npx.set_np()\n",
    "\n",
    "def add(a, b):\n",
    "    return a + b\n",
    "def fancy_func(a, b, c):\n",
    "    e = add(a, b)\n",
    "    return add(c, e)\n",
    "fancy_func(1, 2, 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "符号式编程。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-29T16:56:39.724022Z",
     "start_time": "2019-08-29T16:56:39.362Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n"
     ]
    }
   ],
   "source": [
    "def add_str():\n",
    "    return '''def add(a, b):\n",
    "    return a + b\n",
    "'''\n",
    "def fancy_func_str():\n",
    "    return '''def fancy_func(a, b, c):\n",
    "    e = add(a, b)\n",
    "    return add(c, e)\n",
    "'''\n",
    "def evoke_str():\n",
    "    return add_str() + fancy_func_str() + '''\n",
    "print(fancy_func(1, 2, 3))\n",
    "'''\n",
    "prog = evoke_str()\n",
    "y = compile(prog, '', 'exec')\n",
    "exec(y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "使用 ``HybridSequential`` 类来构建网络。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-29T16:56:39.724910Z",
     "start_time": "2019-08-29T16:56:39.364Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.18370919, -0.02053997]])"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_net():\n",
    "    net = nn.HybridSequential()\n",
    "    net.add(nn.Dense(256, activation='relu'),\n",
    "            nn.Dense(128, activation='relu'),\n",
    "            nn.Dense(2))\n",
    "    net.initialize()\n",
    "    return net\n",
    "\n",
    "x = np.random.normal(size=(1, 512))\n",
    "net = get_net()\n",
    "net(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "编译优化并运行。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-29T16:56:39.725781Z",
     "start_time": "2019-08-29T16:56:39.365Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.18370919, -0.02053997]])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "net.hybridize()\n",
    "net(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "测试速度。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-29T16:56:39.726672Z",
     "start_time": "2019-08-29T16:56:39.366Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "before hybridizing: 0.6084 sec\n",
      "after hybridizing: 0.2034 sec\n"
     ]
    }
   ],
   "source": [
    "def benchmark(net, x):\n",
    "    timer = d2l.Timer()\n",
    "    for i in range(1000):\n",
    "        _ = net(x)\n",
    "    npx.waitall()\n",
    "    return timer.stop()\n",
    "\n",
    "net = get_net()\n",
    "print('before hybridizing: %.4f sec' % (benchmark(net, x)))\n",
    "net.hybridize()\n",
    "print('after hybridizing: %.4f sec' % (benchmark(net, x)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "导出到其他语言。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-29T16:56:39.727577Z",
     "start_time": "2019-08-29T16:56:39.367Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "my_mlp-0000.params  my_mlp-symbol.json\r\n",
      "{\r\n",
      "  \"nodes\": [\r\n",
      "    {\r\n",
      "      \"op\": \"null\", \r\n",
      "      \"name\": \"data\", \r\n",
      "      \"inputs\": []\r\n",
      "    }, \r\n",
      "    {\r\n",
      "      \"op\": \"null\", \r\n",
      "      \"name\": \"dense3_weight\", \r\n",
      "      \"attrs\": {\r\n",
      "        \"__dtype__\": \"0\", \r\n",
      "        \"__lr_mult__\": \"1.0\", \r\n",
      "        \"__shape__\": \"(256, -1)\", \r\n",
      "        \"__storage_type__\": \"0\", \r\n",
      "        \"__wd_mult__\": \"1.0\"\r\n",
      "      }, \r\n",
      "      \"inputs\": []\r\n",
      "    }, \r\n",
      "    {\r\n"
     ]
    }
   ],
   "source": [
    "net.export('my_mlp')\n",
    "!ls my_mlp*\n",
    "!head -n20 my_mlp-symbol.json"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}