{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pASl9O_Gb4x1"
   },
   "source": [
    "# Introduction to Neural Networks\n",
    "\n",
    "Titouan Vayer, titouan.vayer@inria.fr \n",
    "Aurélien Garivier, aurelien.garivier@ens-lyon.fr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "2WZvU5Ocb4x5"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sl9aeSeyb4x7"
   },
   "source": [
    "## Part 1: 1D regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "4ZQ9i0t6b4x8"
   },
   "outputs": [],
   "source": [
    "F = lambda x: 0.5*x**3-3*x**2+1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LlnxgsPJb4x9"
   },
   "source": [
    "### Computing a gradient thanks to autodiff"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XBVf_Uygb4x9"
   },
   "source": [
    "Let $F(x) = 0.5 x^{3} - 3 x^{2} +1$, and it's derivative $F'(x) = \\frac{3}{2} x^{2}- 6 x$.\n",
    "In the context of learning, the function is called by pytorch *loss*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "id": "N1bTZQ7hb4x-"
   },
   "outputs": [],
   "source": [
    "x = torch.Tensor([2])\n",
    "x.requires_grad = True # tell pytorch it needs to compute the gradients wrt x\n",
    "loss = F(x) # computing the function value = forward pass\n",
    "loss.backward() # backward pass, to prepare the computation of the gradient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "r-ZPMonGb4x_",
    "outputId": "d3e2f506-4de4-4f29-f76e-b78dd21ed11d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-7.0"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loss.item() # function value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "bL8uhoCIb4yA",
    "outputId": "4947faf0-1ad3-4e34-9863-244d80c4d0b6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([-6.])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.grad #le gradient en x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vvplZM_Qb4yB"
   },
   "source": [
    "##### Q: Check that the computed values are correct"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "oGjeZamNb4yB"
   },
   "source": [
    "##### Q: Modify the code above to do the same computations at x=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = torch.Tensor([1])\n",
    "x.requires_grad = True # tell pytorch it needs to compute the gradients wrt x\n",
    "loss = F(x) # computing the function value = forward pass\n",
    "loss.backward() # backward pass, to prepare the computation of the gradient"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "yfbWTMhtb4yE"
   },
   "source": [
    "### Simple 1D regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Gevcb_1Nb4yE"
   },
   "source": [
    "We generate random points for training. They follow a simple model \"signal + noise\", where the signal is the function $F$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "id": "nJiYw8oqb4yF"
   },
   "outputs": [],
   "source": [
    "sigma = 50\n",
    "n = 100\n",
    "X = np.linspace(-10,10,n)\n",
    "y = F(X)+sigma*np.random.randn(n,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "id": "dJZ_LvvJb4yF",
    "outputId": "8127e2db-1852-4779-d949-bde6f7a99e5b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X,y,'o',lw=1, label = 'Noisy samples')\n",
    "plt.plot(X, F(X), 'k--', lw = 3, label = 'True F')\n",
    "plt.title('Generated data', fontsize = 14)\n",
    "plt.legend()\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "jE8BekwAb4yG"
   },
   "source": [
    "We use a neural network to learn, from the previously sample points, a regressor that should be as close as possible from the original function $F$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {
    "id": "uFSJhfAsb4yG"
   },
   "outputs": [],
   "source": [
    "class NeuralNetworkRegressor(nn.Module):\n",
    "    def __init__(self, d1 = 10):\n",
    "        super(NeuralNetworkRegressor, self).__init__()\n",
    "        self.flatten = nn.Flatten()\n",
    "        self.linear_relu_stack = nn.Sequential(\n",
    "            nn.Linear(1, d1),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(d1, d1),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(d1, d1),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(d1, d1),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(d1, d1),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(d1, 1),\n",
    "        )\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.flatten(x)\n",
    "        pred = self.linear_relu_stack(x)\n",
    "        return pred"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "z5PKrv42b4yH"
   },
   "source": [
    "#### Q: Comment on the previously chosen network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "isuMhur2b4yH"
   },
   "source": [
    "In order to train our network, we need to \n",
    "  - instantiate its architecture, \n",
    "  - define an optimizer\n",
    "  - choose a learning rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {
    "id": "gPvZQtLIb4yH"
   },
   "outputs": [],
   "source": [
    "model = NeuralNetworkRegressor(50)\n",
    "learning_rate = 1e-2\n",
    "epochs = 500\n",
    "\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
    "Xt = torch.from_numpy(X).type(torch.float).reshape(-1,1) #pytorch manipulates only tensors, not numpy array\n",
    "yt = torch.from_numpy(y).type(torch.float).reshape(-1,1)\n",
    "loss_fn = nn.MSELoss() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {
    "id": "7vfgBQTzb4yI"
   },
   "outputs": [],
   "source": [
    "pytorch_total_params = sum(p.numel() for p in model.parameters())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {
    "id": "xhZGpjgPb4yI",
    "outputId": "c280d159-4b3e-4ae0-d473-c12d5652a1c1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10351\n"
     ]
    }
   ],
   "source": [
    "print(pytorch_total_params) # total number of parameters of the model: comment!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "siV3tSkcb4yI"
   },
   "source": [
    "#### Q: comment on the choice of the loss function"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "V2BAHdQJb4yI"
   },
   "source": [
    "We can train the network by using the following loop:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {
    "id": "0EHQKTbAb4yJ"
   },
   "outputs": [],
   "source": [
    "losses = []\n",
    "for i in range(epochs):\n",
    "    # Compute prediction and loss\n",
    "    pred = model(Xt)\n",
    "    loss = loss_fn(pred, yt)\n",
    "\n",
    "    # Backpropagation\n",
    "    optimizer.zero_grad()\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    losses.append(loss.item())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "YUHNR8z1b4yJ"
   },
   "source": [
    "We can now display the evolution of the training error as a function of the iterations: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {
    "id": "M5ZFnPFAb4yJ",
    "outputId": "932e57e5-1a8c-44bb-e761-4409526f9377"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Iterations')"
      ]
     },
     "execution_count": 176,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.semilogy(losses, lw = 2)\n",
    "plt.title('Training loss', fontsize = 14)\n",
    "plt.xlabel('Iterations', fontsize = 14)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KrPvkrsWb4yK"
   },
   "source": [
    "#### Q: Did the algorithm converge?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "D_ScZSJYb4yK"
   },
   "source": [
    "Let us now show the model prediction, and compare with the original function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "id": "b0uSBzgTb4yK",
    "outputId": "48e152d1-286a-4d9c-ea4a-ac6f62dcc238"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X,y,'o',lw=1)\n",
    "yhat = model(torch.from_numpy(X).type(torch.float).reshape(-1,1)).detach().numpy()\n",
    "plt.plot(X,yhat,'r-',lw=2, label = 'prediction')\n",
    "plt.plot(X, F(X), 'k--', lw = 3, label = 'True F')\n",
    "plt.title('Number of params = {}'.format(pytorch_total_params))\n",
    "plt.legend()\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "bN_lkuWub4yK"
   },
   "source": [
    "#### Q: Comment on the previous experiment. How to validate properly the fitted model?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "0VL4QR0gb4yL"
   },
   "source": [
    "#### Q: What happens when the number of parameters explodes ? (for example, take $d_1 = 200$)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {
    "id": "2rALieR2b4yL"
   },
   "outputs": [],
   "source": [
    "model2 = NeuralNetworkRegressor(200)\n",
    "learning_rate = 5e-2\n",
    "epochs = 200\n",
    "\n",
    "optimizer = torch.optim.Adam(model2.parameters(), lr=learning_rate)\n",
    "Xt = torch.from_numpy(X).type(torch.float).reshape(-1,1) \n",
    "yt = torch.from_numpy(y).type(torch.float).reshape(-1,1)\n",
    "loss_fn = nn.MSELoss() \n",
    "pytorch_total_params2 = sum(p.numel() for p in model2.parameters())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {
    "id": "0KoMOqznb4yM"
   },
   "outputs": [],
   "source": [
    "losses = []\n",
    "for i in range(epochs):\n",
    "    # Compute prediction and loss\n",
    "    pred = model2(Xt)\n",
    "    loss = loss_fn(pred, yt)\n",
    "\n",
    "    # Backpropagation\n",
    "    optimizer.zero_grad()\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    losses.append(loss.item())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {
    "id": "dqkEbOPcb4yM",
    "outputId": "2c64a594-1b8d-4033-d31c-1fce7d99118e"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Iterations')"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.semilogy(losses, lw = 2)\n",
    "plt.title('Training loss', fontsize = 14)\n",
    "plt.xlabel('Iterations', fontsize = 14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {
    "id": "guQ3XGtvb4yM",
    "outputId": "5e58a7fa-6a5b-4c2f-a38d-e04ab0721ed3"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X,y,'o',lw=1)\n",
    "yhat = model2(torch.from_numpy(X).type(torch.float).reshape(-1,1)).detach().numpy()\n",
    "plt.plot(X,yhat,'r-',lw=2, label = 'prediction')\n",
    "plt.plot(X, F(X), 'k--', lw = 3, label = 'True F')\n",
    "plt.title('Number of params = {}'.format(pytorch_total_params2))\n",
    "plt.legend()\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "provenance": []
  },
  "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}
