03_Pytorch/02_xavier&he.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Xavier Initialization vs He Initialization\n",
    "## 1. 개요\n",
    "- Initialization 의 방법에 따라, 학습률과 학습된 모델의 성능을 비교해보고자 합니다.\n",
    "- MNIST data 를 이용하여, 같은 깊이와 노드 수를 가진 신경망이 Initialization 에 따라 어떤 차이를 보이는지 실험해 보겠습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Import Packages and Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "import torch.nn.functional as F\n",
    "\n",
    "import torchvision.datasets as datasets\n",
    "import torchvision.transforms as transforms\n",
    "\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "mnist_train = datasets.MNIST(root=\"data/\", train=True, transform=transforms.ToTensor(), download=True)\n",
    "mnist_test = datasets.MNIST(root=\"data/\", train=False, transform=transforms.ToTensor(), download=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Modeling\n",
    "\n",
    "- 비교해볼 모델은 다음과 같습니다.    \n",
    "1. Weight Initialization 을 해주지 않은 Linear\n",
    "    - Initialization 을 해주지 않았을 때 pytorch는 아래코드와 같이 uniform distribution 으로 initialize를 하는 것을 확인할 수 있습니다.\n",
    "2. Xavier Normal Initialization 을 해준 Linear\n",
    "3. Xavier Uniform Initialization 을 해준 Linear\n",
    "4. He Normal Initialization 을 해준 Linear\n",
    "5. He Uniform Initialization 을 해준 Linear\n",
    "\n",
    "\n",
    "```python\n",
    "def reset_parameters(self):\n",
    "    stdv = 1. / math.sqrt(self.weight.size(1))\n",
    "    self.weight.data.uniform_(-stdv, stdv)\n",
    "    if self.bias is not None:\n",
    "        self.bias.data.uniform_(-stdv, stdv)\n",
    "```\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "class NoInitialize(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.linear1 = nn.Linear(784, 256, bias=True)\n",
    "        self.linear2 = nn.Linear(256, 128, bias=True)\n",
    "        self.linear3 = nn.Linear(128, 10, bias=True)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        output1 = torch.relu(self.linear1(x))\n",
    "        output2 = torch.relu(self.linear2(output1))\n",
    "        output3 = self.linear3(output2)\n",
    "        return output3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class XavierNormal(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.linear1 = nn.Linear(784, 256, bias=True)\n",
    "        self.linear2 = nn.Linear(256, 128, bias=True)\n",
    "        self.linear3 = nn.Linear(128, 10, bias=True)\n",
    "        \n",
    "        nn.init.xavier_normal_(self.linear1.weight)\n",
    "        nn.init.xavier_normal_(self.linear2.weight)\n",
    "        nn.init.xavier_normal_(self.linear3.weight)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        output1 = torch.relu(self.linear1(x))\n",
    "        output2 = torch.relu(self.linear2(output1))\n",
    "        output3 = self.linear3(output2)\n",
    "        return output3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class XavierUniform(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.linear1 = nn.Linear(784, 256, bias=True)\n",
    "        self.linear2 = nn.Linear(256, 128, bias=True)\n",
    "        self.linear3 = nn.Linear(128, 10, bias=True)\n",
    "        \n",
    "        nn.init.xavier_uniform_(self.linear1.weight)\n",
    "        nn.init.xavier_uniform_(self.linear2.weight)\n",
    "        nn.init.xavier_uniform_(self.linear3.weight)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        output1 = torch.relu(self.linear1(x))\n",
    "        output2 = torch.relu(self.linear2(output1))\n",
    "        output3 = self.linear3(output2)\n",
    "        return output3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "class HeNormal(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.linear1 = nn.Linear(784, 256, bias=True)\n",
    "        self.linear2 = nn.Linear(256, 128, bias=True)\n",
    "        self.linear3 = nn.Linear(128, 10, bias=True)\n",
    "        \n",
    "        nn.init.kaiming_normal_(self.linear1.weight)\n",
    "        nn.init.kaiming_normal_(self.linear2.weight)\n",
    "        nn.init.kaiming_normal_(self.linear3.weight)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        output1 = torch.relu(self.linear1(x))\n",
    "        output2 = torch.relu(self.linear2(output1))\n",
    "        output3 = self.linear3(output2)\n",
    "        return output3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "class HeUniform(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.linear1 = nn.Linear(784, 256, bias=True)\n",
    "        self.linear2 = nn.Linear(256, 128, bias=True)\n",
    "        self.linear3 = nn.Linear(128, 10, bias=True)\n",
    "        \n",
    "        nn.init.kaiming_uniform_(self.linear1.weight)\n",
    "        nn.init.kaiming_uniform_(self.linear2.weight)\n",
    "        nn.init.kaiming_uniform_(self.linear3.weight)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        output1 = torch.relu(self.linear1(x))\n",
    "        output2 = torch.relu(self.linear2(output1))\n",
    "        output3 = self.linear3(output2)\n",
    "        return output3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[NoInitialize(\n",
       "   (linear1): Linear(in_features=784, out_features=256, bias=True)\n",
       "   (linear2): Linear(in_features=256, out_features=128, bias=True)\n",
       "   (linear3): Linear(in_features=128, out_features=10, bias=True)\n",
       " ), XavierNormal(\n",
       "   (linear1): Linear(in_features=784, out_features=256, bias=True)\n",
       "   (linear2): Linear(in_features=256, out_features=128, bias=True)\n",
       "   (linear3): Linear(in_features=128, out_features=10, bias=True)\n",
       " ), XavierUniform(\n",
       "   (linear1): Linear(in_features=784, out_features=256, bias=True)\n",
       "   (linear2): Linear(in_features=256, out_features=128, bias=True)\n",
       "   (linear3): Linear(in_features=128, out_features=10, bias=True)\n",
       " ), HeNormal(\n",
       "   (linear1): Linear(in_features=784, out_features=256, bias=True)\n",
       "   (linear2): Linear(in_features=256, out_features=128, bias=True)\n",
       "   (linear3): Linear(in_features=128, out_features=10, bias=True)\n",
       " ), HeUniform(\n",
       "   (linear1): Linear(in_features=784, out_features=256, bias=True)\n",
       "   (linear2): Linear(in_features=256, out_features=128, bias=True)\n",
       "   (linear3): Linear(in_features=128, out_features=10, bias=True)\n",
       " )]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models = []\n",
    "\n",
    "no_initial = NoInitialize()\n",
    "xavier_normal = XavierNormal()\n",
    "xavier_uniform = XavierUniform()\n",
    "he_normal = HeNormal()\n",
    "he_uniform = HeUniform()\n",
    "\n",
    "models.append(no_initial)\n",
    "models.append(xavier_normal)\n",
    "models.append(xavier_uniform)\n",
    "models.append(he_normal)\n",
    "models.append(he_uniform)\n",
    "models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NoInitialize  - epoch :  tensor(0.0421, grad_fn=<AddBackward0>)\n",
      "NoInitialize  - epoch :  tensor(0.0155, grad_fn=<AddBackward0>)\n",
      "NoInitialize  - epoch :  tensor(0.0088, grad_fn=<AddBackward0>)\n",
      "XavierNormal  - epoch :  tensor(0.0338, grad_fn=<AddBackward0>)\n",
      "XavierNormal  - epoch :  tensor(0.0126, grad_fn=<AddBackward0>)\n",
      "XavierNormal  - epoch :  tensor(0.0076, grad_fn=<AddBackward0>)\n",
      "XavierUniform  - epoch :  tensor(0.0358, grad_fn=<AddBackward0>)\n",
      "XavierUniform  - epoch :  tensor(0.0148, grad_fn=<AddBackward0>)\n",
      "XavierUniform  - epoch :  tensor(0.0102, grad_fn=<AddBackward0>)\n",
      "HeNormal  - epoch :  tensor(0.0333, grad_fn=<AddBackward0>)\n",
      "HeNormal  - epoch :  tensor(0.0134, grad_fn=<AddBackward0>)\n",
      "HeNormal  - epoch :  tensor(0.0064, grad_fn=<AddBackward0>)\n",
      "HeUniform  - epoch :  tensor(0.0328, grad_fn=<AddBackward0>)\n",
      "HeUniform  - epoch :  tensor(0.0136, grad_fn=<AddBackward0>)\n",
      "HeUniform  - epoch :  tensor(0.0094, grad_fn=<AddBackward0>)\n",
      "TRAINING DONE\n"
     ]
    }
   ],
   "source": [
    "batch_size = 100\n",
    "lr = 0.001\n",
    "epochs = 15\n",
    "\n",
    "dataloader = DataLoader(dataset=mnist_train, batch_size=batch_size, shuffle=True, drop_last=True)\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "total_batch = len(dataloader)\n",
    "\n",
    "model_losses = []\n",
    "\n",
    "for model in models:\n",
    "    optimizer = optim.Adam(model.parameters(), lr=lr)\n",
    "    total_loss = []\n",
    "    for epoch in range(1, epochs + 1):\n",
    "        \n",
    "        avg_cost = 0\n",
    "        for x_batch, y_batch in dataloader:\n",
    "            x_batch = x_batch.view(-1, 784)\n",
    "            y_hat = model(x_batch)\n",
    "            loss = criterion(y_hat, y_batch)\n",
    "            \n",
    "            optimizer.zero_grad()\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "            \n",
    "            avg_cost += loss / total_batch\n",
    "        total_loss.append(avg_cost)\n",
    "        if epoch % 5 == 0:\n",
    "            print(model.__class__.__name__, ' - epoch : ', avg_cost)    \n",
    "    model_losses.append(total_loss)\n",
    "    \n",
    "print(\"TRAINING DONE\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Compare Training Loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 500,
       "width": 625
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 8))\n",
    "for model, model_loss in list(zip(models, model_losses)):\n",
    "    plt.plot(model_loss, label=model.__class__.__name__, linewidth=2)\n",
    "\n",
    "plt.legend()\n",
    "plt.xlabel(\"epoch\")\n",
    "plt.ylabel(\"training loss\")\n",
    "plt.title(\"Training loss by Initialization \")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Compare Test Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NoInitialize  - Accuracy: 0.9787999987602234\n",
      "XavierNormal  - Accuracy: 0.980400025844574\n",
      "XavierUniform  - Accuracy: 0.9803000092506409\n",
      "HeNormal  - Accuracy: 0.974399983882904\n",
      "HeUniform  - Accuracy: 0.9782999753952026\n"
     ]
    }
   ],
   "source": [
    "with torch.no_grad():\n",
    "    x_test = mnist_test.test_data.view(-1, 784).float()\n",
    "    y_test = mnist_test.test_labels\n",
    "\n",
    "    for model in models:\n",
    "        prediction = model(x_test)\n",
    "        correct_prediction = torch.argmax(prediction, 1) == y_test\n",
    "        accuracy = correct_prediction.float().mean()\n",
    "        print(model.__class__.__name__, ' - Accuracy:', accuracy.item())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. Conclusion\n",
    "- 먼저, Training Loss 가 떨어지는 것을 보면, Default Initialization 보다, 나머지 Initialization 방법에서 더 좋은 성과가 있음을 확인 할 수 있습니다. Epoch 가 진행됨에 따라 더 안정적이고 빠르게 training loss 가 감소 하는 것을 볼 수 있습니다.\n",
    "- 또한 Test set 에서의 accuracy 를 통해 Xavier Initialization 방법이 성능역시 뛰어난 것을 확인 할 수 있습니다. 이는 같은 epoch 동안에 training 시켰을 때, 초기화 하는 방법에 따라 학습률이 다르다는 것을 의미하는 것이기도 합니다.\n",
    "- 하지만 이것이 절대적인 초기화 방법의 유리함을 의미하는 것이진 않습니다. 초기화 방법 역시, problemmatic 한 parameter의 한 종류 이므로, 문제마다 다양하게 시도해 봐야 합니다."
   ]
  }
 ],
 "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}