02_20_2023 Linear regression using pytorch.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b930f522",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "from torch.utils import data\n",
    "from d2l import torch as d2l\n",
    "\n",
    "true_w = torch.tensor([2, -3.4])\n",
    "true_b = 4.2\n",
    "features, labels = d2l.synthetic_data(true_w, true_b, 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "9229485e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_array(data_arrays, batch_size, is_train=True):  #@save\n",
    "    \"\"\"构造一个PyTorch数据迭代器\"\"\"\n",
    "    dataset = data.TensorDataset(*data_arrays)\n",
    "    return data.DataLoader(dataset, batch_size, shuffle=is_train)\n",
    "\n",
    "batch_size = 10\n",
    "data_iter = load_array((features, labels), batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "3c06bab3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[tensor([[-0.2233, -0.4359],\n",
       "         [-0.4633,  0.4167],\n",
       "         [-0.0104,  0.8766],\n",
       "         [-0.6802,  0.2990],\n",
       "         [ 0.3793, -0.2224],\n",
       "         [-1.2978, -0.5319],\n",
       "         [-0.2682, -0.1509],\n",
       "         [-0.0739, -0.7883],\n",
       "         [-0.5356, -0.7860],\n",
       "         [ 1.6555,  2.2472]]),\n",
       " tensor([[ 5.2460],\n",
       "         [ 1.8579],\n",
       "         [ 1.1977],\n",
       "         [ 1.8342],\n",
       "         [ 5.7152],\n",
       "         [ 3.4234],\n",
       "         [ 4.1884],\n",
       "         [ 6.7469],\n",
       "         [ 5.8020],\n",
       "         [-0.1241]])]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "next(iter(data_iter))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "957027c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# nn是神经网络的缩写\n",
    "from torch import nn\n",
    "\n",
    "net = nn.Sequential(nn.Linear(2, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "39c66ab3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([0.])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "net[0].weight.data.normal_(0, 0.01)\n",
    "net[0].bias.data.fill_(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "8dacd288",
   "metadata": {},
   "outputs": [],
   "source": [
    "loss = nn.MSELoss()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "aac3594b",
   "metadata": {},
   "outputs": [],
   "source": [
    "trainer = torch.optim.SGD(net.parameters(), lr=0.03)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a79f2097",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 1, loss 0.000328\n",
      "epoch 2, loss 0.000101\n",
      "epoch 3, loss 0.000101\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 3\n",
    "for epoch in range(num_epochs):\n",
    "    for X, y in data_iter:\n",
    "        l = loss(net(X) ,y)\n",
    "        trainer.zero_grad()\n",
    "        l.backward()\n",
    "        trainer.step()\n",
    "    l = loss(net(features), labels)\n",
    "    print(f'epoch {epoch + 1}, loss {l:f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "1bacdeca",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "w的估计误差： tensor([-0.0007, -0.0002])\n",
      "b的估计误差： tensor([-0.0001])\n"
     ]
    }
   ],
   "source": [
    "w = net[0].weight.data\n",
    "print('w的估计误差：', true_w - w.reshape(true_w.shape))\n",
    "b = net[0].bias.data\n",
    "print('b的估计误差：', true_b - b)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "659561aa",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}