learn_pytorch_23_pytorch实战项目之一元线性回归2完整版.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 1 导入需要的模块和包 \n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt \n",
    "import torch \n",
    "# 嵌入模式\n",
    "%matplotlib inline   \n",
    "\n",
    "#  2 生成数据集  \n",
    "# 2.1 特征数据 \n",
    "x_data = np.linspace(1,3,num=50,endpoint=True,retstep=0,dtype=np.float32)\n",
    "\n",
    "# 2.2 目标数据：添加噪声 \n",
    "diff = np.random.randn(50)\n",
    "y_data = 3 * x_data+0.1 + diff\n",
    "y_data = np.array(y_data, dtype=np.float32) \n",
    "\n",
    "# 2.3 组建数据集:将所有样本组合起来 \n",
    "x_data = np.expand_dims(x_data,axis=1)\n",
    "y_data = np.expand_dims(y_data,axis=1)\n",
    "\n",
    "# 2.4 画图展示结果 \n",
    "plt.plot(x_data,y_data,\"ro\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "# 3 创建线性回归模型 \n",
    "class MySimpleLinearRegression(torch.nn.Module):\n",
    "    def __init__(self):\n",
    "        super(MySimpleLinearRegression,self).__init__()\n",
    "        self.linear = torch.nn.Linear(1,1)\n",
    "    \n",
    "    def forward(self,x):\n",
    "        out = self.linear(x)\n",
    "        return out \n",
    "\n",
    "    \n",
    "# 4 训练模型并测试\n",
    "def train_and_test():\n",
    "    # 4.1 创建模型，选择GPU或者cpu\n",
    "    if torch.cuda.is_available():\n",
    "        device = torch.device(\"cuda:0\")\n",
    "        model = MySimpleLinearRegression()\n",
    "        model.to(device)\n",
    "    else:\n",
    "        device = torch.device(\"cpu\")\n",
    "        model = MySimpleLinearRegression()\n",
    "        model.to(device)\n",
    "        \n",
    "    # 4.2 创建优化目标和优化方法\n",
    "    \n",
    "    criterion = torch.nn.MSELoss() # 均方误差 \n",
    "    optimizer = torch.optim.SGD(model.parameters(),lr=0.01) # 随机梯度下降优化\n",
    "    \n",
    "    for i in range(500):\n",
    "        i+=1\n",
    "        \n",
    "        # numpy数据转为tensor \n",
    "        inputs = torch.from_numpy(x_data).to(device)\n",
    "        labels = torch.from_numpy(y_data).to(device)\n",
    "        \n",
    "        # 4.3 前向传播 \n",
    "        output = model(inputs)\n",
    "        \n",
    "        # 4.4 计算损失\n",
    "        loss = criterion(output, labels)\n",
    "        \n",
    "        # 4.5 反向传播\n",
    "        # 梯度要清0\n",
    "        optimizer.zero_grad()\n",
    "        \n",
    "        loss.backward()\n",
    "        \n",
    "        # 更新权重参数\n",
    "        optimizer.step()\n",
    "\n",
    "        # 打印\n",
    "        if i % 10 == 0:\n",
    "            print('loss值：', loss.item())\n",
    "        \n",
    "    # 4.6 测试模式\n",
    "    model.eval()\n",
    "    y_predict = model(torch.from_numpy(x_data)).data.numpy()\n",
    "    plt.plot(x_data, y_data, \"ro\")\n",
    "    plt.plot(x_data, y_predict, \"k\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss值： 6.654595851898193\n",
      "loss值： 2.0426411628723145\n",
      "loss值： 1.5370522737503052\n",
      "loss值： 1.4721521139144897\n",
      "loss值： 1.4547921419143677\n",
      "loss值： 1.442800760269165\n",
      "loss值： 1.4316481351852417\n",
      "loss值： 1.4208414554595947\n",
      "loss值： 1.4103214740753174\n",
      "loss值： 1.40007483959198\n",
      "loss值： 1.390094518661499\n",
      "loss值： 1.3803731203079224\n",
      "loss值： 1.3709038496017456\n",
      "loss值： 1.3616803884506226\n",
      "loss值： 1.352696418762207\n",
      "loss值： 1.3439455032348633\n",
      "loss值： 1.3354216814041138\n",
      "loss值： 1.3271191120147705\n",
      "loss值： 1.3190319538116455\n",
      "loss值： 1.311154842376709\n",
      "loss值： 1.3034818172454834\n",
      "loss值： 1.2960079908370972\n",
      "loss值： 1.2887282371520996\n",
      "loss值： 1.2816375494003296\n",
      "loss值： 1.2747308015823364\n",
      "loss值： 1.268003225326538\n",
      "loss值： 1.261449933052063\n",
      "loss值： 1.255067229270935\n",
      "loss值： 1.248849868774414\n",
      "loss值： 1.2427939176559448\n",
      "loss值： 1.236895203590393\n",
      "loss值： 1.2311495542526245\n",
      "loss值： 1.225553035736084\n",
      "loss值： 1.2201017141342163\n",
      "loss值： 1.2147917747497559\n",
      "loss值： 1.2096195220947266\n",
      "loss值： 1.2045819759368896\n",
      "loss值： 1.1996747255325317\n",
      "loss值： 1.1948949098587036\n",
      "loss值： 1.190239429473877\n",
      "loss值： 1.1857043504714966\n",
      "loss值： 1.1812872886657715\n",
      "loss值： 1.1769845485687256\n",
      "loss值： 1.1727936267852783\n",
      "loss值： 1.168711543083191\n",
      "loss值： 1.1647354364395142\n",
      "loss值： 1.1608624458312988\n",
      "loss值： 1.1570900678634644\n",
      "loss值： 1.1534154415130615\n",
      "loss值： 1.1498360633850098\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_and_test()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}