hw13_meta_regression.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "hw13_meta_regression.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Iallen520/lhy_DL_Hw/blob/master/hw13_meta_regression.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yow9nZCqtvjj",
        "colab_type": "text"
      },
      "source": [
        "# **Homework 13 - Meta Learning**\n",
        "\n",
        "若有任何問題，歡迎來信至助教信箱 ntu-ml-2020spring-ta@googlegroups.com"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SWbkcznitlaF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.utils.data as data\n",
        "import torch.nn.functional as F\n",
        "import numpy as np\n",
        "from tqdm import tqdm\n",
        "import copy\n",
        "import matplotlib.pyplot as plt\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LSGlyeWsxTDI",
        "colab_type": "text"
      },
      "source": [
        "生成 $ a*\\sin(x+b) $ 的資料點，其中$a, b$的範圍分別預設為$[0.1, 5], [0, 2\\pi]$，每一個 $ a*\\sin(x+b) $的函數有10個資料點當作訓練資料。測試時則可用較密集的資料點直接由畫圖來看generalize的好壞。\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G4tjWphivAUD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "device = 'cpu'\n",
        "def meta_task_data(seed = 0, a_range=[0.1, 5], b_range = [0, 2*np.pi], task_num = 100,\n",
        "                   n_sample = 10, sample_range = [-5, 5], plot = False):\n",
        "    np.random.seed = seed\n",
        "    a_s = np.random.uniform(low = a_range[0], high = a_range[1], size = task_num)\n",
        "    b_s = np.random.uniform(low = b_range[0], high = b_range[1], size = task_num)\n",
        "    total_x = []\n",
        "    total_y = []\n",
        "    label = []\n",
        "    for t in range(task_num):\n",
        "        x = np.random.uniform(low = sample_range[0], high = sample_range[1], size = n_sample)\n",
        "        total_x.append(x)\n",
        "        total_y.append( a_s[t]*np.sin(x+b_s[t]) )\n",
        "        label.append('{:.3}*sin(x+{:.3})'.format(a_s[t], b_s[t]))\n",
        "    if plot:\n",
        "        plot_x = [np.linspace(-5, 5, 1000)]\n",
        "        plot_y = []\n",
        "        for t in range(task_num):\n",
        "            plot_y.append( a_s[t]*np.sin(plot_x+b_s[t]) ) \n",
        "        return total_x, total_y, plot_x, plot_y, label\n",
        "    else:\n",
        "        return total_x, total_y, label"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zjKn7O5H0WMA",
        "colab_type": "text"
      },
      "source": [
        "以下我們將老師MAML投影片第27頁的$\\phi$稱作meta weight，$\\theta$則稱為sub weight。\n",
        "\n",
        "老師投影片： http://speech.ee.ntu.edu.tw/~tlkagk/courses/ML_2019/Lecture/Meta1%20(v6).pdf\n",
        "\n",
        "***\n",
        "\n",
        "為了讓sub weight的gradient能夠傳到meta weight  (因為sub weight的初始化是從meta weight來的，所以想當然我們用sub weight算出來的loss對meta weight也應該是可以算gradient才對)，這邊我們需要重新定義一些pytorch內的layer的運算。\n",
        "\n",
        "實際上*MetaLinear*這個class做的事情跟torch.nn.Linear完全是一樣的，唯一的差別在於這邊的每一個tensor都沒有被變成torch.nn.Parameter。這麼做的原因是因為等一下我們從meta weight那裏複製(init weight輸入meta weight後weight與bias使用.clone)的時候，tensor的clone的操作是可以傳遞gradient的，以方便我們用gradient更新meta weight。這個寫法的代價是我們就沒辦法使用torch.optim更新sub weight了，因為參數都只用tensor紀錄。也因此我們接下來需要自己寫gradient update的函數(只用SGD的話是簡單的)。"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HCagZZYPwr6p",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class MetaLinear(nn.Module):\n",
        "    def __init__(self, init_layer = None):\n",
        "        super(MetaLinear, self).__init__()\n",
        "        if type(init_layer) != type(None):\n",
        "            self.weight = init_layer.weight.clone()\n",
        "            self.bias = init_layer.bias.clone()\n",
        "    def zero_grad(self):\n",
        "        self.weight.grad  = torch.zeros_like(self.weight)\n",
        "        self.bias.grad  = torch.zeros_like(self.bias)\n",
        "    def forward(self, x):\n",
        "        return F.linear(x, self.weight, self.bias)\n",
        "        \n",
        "\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LJBv53j53xIp",
        "colab_type": "text"
      },
      "source": [
        "這裡的forward和一般的model是一樣的，唯一的差別在於我們需要多寫一下\\_\\_init\\_\\_函數讓他比起一般的pytorch model多一個可以從meta weight複製的功能(這邊因為我把model的架構寫死了所以可能看起來會有點多餘，讀者可以自己把net()改成可以自己調整架構的樣子，然後思考一下如何生成一個跟meta weight一樣形狀的sub weight)\n",
        "\n",
        "update函數就如同前一段提到的，我們需要自己先手動用SGD更新一次sub weight，接著再使用下一步的gradient(第二步)來更新meta weight。zero_grad函數在此處沒有用到，因為實際上我們計算第二步的gradient時會需要第一步的grad，這也是為什麼我們第一次backward的時候需要create_graph=True  (建立計算圖以計算二階的gradient)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FdInqRZ93oiG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class net(nn.Module):\n",
        "    def __init__(self, init_weight=None):\n",
        "        super(net, self).__init__()\n",
        "        if type(init_weight) != type(None):\n",
        "            for name, module in init_weight.named_modules():\n",
        "                if name != '':\n",
        "                    setattr(self, name, MetaLinear(module))\n",
        "        else:\n",
        "            self.hidden1 = nn.Linear(1, 40)\n",
        "            self.hidden2 = nn.Linear(40, 40)\n",
        "            self.out = nn.Linear(40, 1)\n",
        "    \n",
        "    def zero_grad(self):\n",
        "        layers = self.__dict__['_modules']\n",
        "        for layer in layers.keys():\n",
        "            layers[layer].zero_grad()\n",
        "    def update(self, parent, lr = 1):\n",
        "        layers = self.__dict__['_modules']\n",
        "        parent_layers = parent.__dict__['_modules']\n",
        "        for param in layers.keys():\n",
        "            layers[param].weight = layers[param].weight - lr*parent_layers[param].weight.grad\n",
        "            layers[param].bias = layers[param].bias - lr*parent_layers[param].bias.grad\n",
        "        # gradient will flow back due to clone backward\n",
        "        \n",
        "    def forward(self, x):\n",
        "        x = F.relu(self.hidden1(x))\n",
        "        x = F.relu(self.hidden2(x))\n",
        "        return self.out(x)\n",
        "\n",
        "    \n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LB06F2Fy7gHc",
        "colab_type": "text"
      },
      "source": [
        "前面的class中我們已經都將複製meta weight到sub weight，以及sub weight的更新，gradient的傳遞都搞定了，meta weight自己本身的參數就可以按照一般pytorch model的模式，使用torch.optim來更新了。\n",
        "\n",
        "gen_model函數做的事情其實就是產生N個sub weight，並且使用前面我們寫好的複製meta weight的功能。\n",
        "\n",
        "注意到複製weight其實是整個code的關鍵，因為我們需要將sub weight計算的第二步gradient正確的傳回meta weight。讀者從meta weight與sub weight更新參數作法的差別(手動更新/用torch.nn.Parameter與torch.optim)可以再思考一下兩者的差別。\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BwwEtUXU7fm0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Meta_learning_model():\n",
        "    def __init__(self, init_weight = None):\n",
        "        super(Meta_learning_model, self).__init__()\n",
        "        self.model = net().to(device)\n",
        "        if type(init_weight) != type(None):\n",
        "            self.model.load_state_dict(init_weight)\n",
        "        self.grad_buffer = 0\n",
        "    def gen_models(self, num, check = True):\n",
        "        models = [net(init_weight=self.model).to(device) for i in range(num)]\n",
        "        return models\n",
        "    def clear_buffer(self):\n",
        "        print(\"Before grad\", self.grad_buffer)\n",
        "        self.grad_buffer = 0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jpxeb_ew8qb-",
        "colab_type": "text"
      },
      "source": [
        "接下來就是生成訓練/測試資料，建立meta weightmeta weight的模型以及用來比較的model pretraining的模型"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b376YYA1vF9a",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "bsz = 10\n",
        "train_x, train_y, train_label = meta_task_data(task_num=50000*10) \n",
        "train_x = torch.Tensor(train_x).unsqueeze(-1) # add one dim\n",
        "train_y = torch.Tensor(train_y).unsqueeze(-1)\n",
        "train_dataset = data.TensorDataset(train_x, train_y)\n",
        "train_loader = data.DataLoader(dataset=train_dataset, batch_size=bsz, shuffle=False)\n",
        "\n",
        "test_x, test_y, plot_x, plot_y, test_label = meta_task_data(task_num=1, n_sample = 10, plot=True)  \n",
        "test_x = torch.Tensor(test_x).unsqueeze(-1) # add one dim\n",
        "test_y = torch.Tensor(test_y).unsqueeze(-1) # add one dim\n",
        "plot_x = torch.Tensor(plot_x).unsqueeze(-1) # add one dim\n",
        "test_dataset = data.TensorDataset(test_x, test_y)\n",
        "test_loader = data.DataLoader(dataset=test_dataset, batch_size=bsz, shuffle=False)  \n",
        "\n",
        "\n",
        "meta_model = Meta_learning_model()\n",
        "\n",
        "meta_optimizer = torch.optim.Adam(meta_model.model.parameters(), lr = 1e-3)\n",
        "\n",
        "\n",
        "pretrain = net()\n",
        "pretrain.to(device)\n",
        "pretrain.train()\n",
        "pretrain_optim = torch.optim.Adam(pretrain.parameters(), lr = 1e-3)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cOVC8J2z9UUO",
        "colab_type": "text"
      },
      "source": [
        "進行訓練，注意一開始我們要先生成一群sub weight(code裡面的sub models)，然後將一個batch的不同的sin函數的10筆資料點拿來訓練sub weight。注意這邊sub weight計算第一步gradient與第二步gradient時使用各五筆不重複的資料點(因此使用[:5]與[5:]來取)。但在訓練model pretraining的對照組時則沒有這個問題(所以pretraining的model是可以確實的走兩步gradient的)\n",
        "\n",
        "每一個sub weight計算完loss後相加(內層的for迴圈)後就可以使用optimizer來更新meta weight，再次提醒一下sub weight計算第一次loss的時候backward是需要create_graph=True的，這樣計算第二步gradient的時候才會真的計算到二階的項。讀者可以在這個地方思考一下如何將這段程式碼改成MAML的一階做法。"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wyC2Mv86vP8l",
        "colab_type": "code",
        "outputId": "c5b2310f-fc69-435e-a0c2-26eb51fd1ddc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "epoch = 1\n",
        "for e in range(epoch):\n",
        "    meta_model.model.train()\n",
        "    for x, y in tqdm(train_loader):\n",
        "        x = x.to(device)\n",
        "        y = y.to(device)\n",
        "        sub_models = meta_model.gen_models(bsz)\n",
        "\n",
        "        meta_l = 0\n",
        "        for model_num in range(len(sub_models)):\n",
        "            \n",
        "            sample = torch.randint(0,10, size = (10,), dtype = torch.long)\n",
        "            \n",
        "            #pretraining\n",
        "            pretrain_optim.zero_grad()\n",
        "            y_tilde = pretrain(x[model_num][sample[:5],:])\n",
        "            little_l = F.mse_loss(y_tilde, y[model_num][sample[:5],:])\n",
        "            little_l.backward()\n",
        "            pretrain_optim.step()\n",
        "            pretrain_optim.zero_grad()\n",
        "            y_tilde = pretrain(x[model_num][sample[5:],:])\n",
        "            little_l = F.mse_loss(y_tilde, y[model_num][sample[5:],:])\n",
        "            little_l.backward()\n",
        "            pretrain_optim.step()\n",
        "            \n",
        "            # meta learning\n",
        "            \n",
        "            y_tilde = sub_models[model_num](x[model_num][sample[:5],:])\n",
        "            little_l = F.mse_loss(y_tilde, y[model_num][sample[:5],:])\n",
        "            #計算第一次gradient並保留計算圖以接著計算更高階的gradient\n",
        "            little_l.backward(create_graph = True)\n",
        "            sub_models[model_num].update(lr = 1e-2, parent = meta_model.model)\n",
        "            #先清空optimizer中計算的gradient值(避免累加)\n",
        "            meta_optimizer.zero_grad()\n",
        "            \n",
        "            #計算第二次(二階)的gradient，二階的原因來自第一次update時有計算過一次gradient了\n",
        "            y_tilde = sub_models[model_num](x[model_num][sample[5:],:])\n",
        "            meta_l =  meta_l + F.mse_loss(y_tilde, y[model_num][sample[5:],:])\n",
        "\n",
        "        meta_l = meta_l / bsz\n",
        "        meta_l.backward()\n",
        "        meta_optimizer.step()\n",
        "        meta_optimizer.zero_grad()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 50000/50000 [35:45<00:00, 22.86it/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5SQ4g_qA_gtr",
        "colab_type": "text"
      },
      "source": [
        "測試我們訓練好的meta weight"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iHr-BPWEv9W_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test_model = copy.deepcopy(meta_model.model)\n",
        "test_model.train()\n",
        "test_optim = torch.optim.SGD(test_model.parameters(), lr = 1e-3)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KccSKypP_nMT",
        "colab_type": "text"
      },
      "source": [
        "先畫出待測試的sin函數，以及用圓點點出測試時給meta weight訓練的十筆資料點"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "abeYF21Ev_tU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "fig = plt.figure(figsize = [9.6,7.2])\n",
        "ax = plt.subplot(111)\n",
        "plot_x1 = plot_x.squeeze().numpy()\n",
        "ax.scatter(test_x.numpy().squeeze(), test_y.numpy().squeeze())\n",
        "ax.plot(plot_x1, plot_y[0].squeeze())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P6THCM8G_0pM",
        "colab_type": "text"
      },
      "source": [
        "分別利用十筆資料點更新meta weight以及pretrained model一個step"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ji6B2UuewDgt",
        "colab_type": "code",
        "outputId": "ee1cbc41-d9b8-4b44-85f8-351bcf4cbc95",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "source": [
        "test_model.train()\n",
        "pretrain.train()\n",
        "\n",
        "for epoch in range(1):\n",
        "    for x, y in test_loader:\n",
        "        y_tilde = test_model(x[0])\n",
        "        little_l = F.mse_loss(y_tilde, y[0])\n",
        "        test_optim.zero_grad()\n",
        "        little_l.backward()\n",
        "        test_optim.step()\n",
        "        print(\"(meta)))Loss: \", little_l.item())\n",
        "\n",
        "for epoch in range(1):\n",
        "    for x, y in test_loader:\n",
        "        y_tilde = pretrain(x[0])\n",
        "        little_l = F.mse_loss(y_tilde, y[0])\n",
        "        pretrain_optim.zero_grad()\n",
        "        little_l.backward()\n",
        "        pretrain_optim.step()\n",
        "        print(\"(pretrain)Loss: \", little_l.item())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(meta)))Loss:  0.3323650658130646\n",
            "(pretrain)Loss:  2.0638670921325684\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NVp1LUdVABZH",
        "colab_type": "text"
      },
      "source": [
        "將更新後的模型所代表的函數繪製出來，與真實的sin函數比較"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-74T_CrXwMJ_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test_model.eval()\n",
        "pretrain.eval()\n",
        "\n",
        "plot_y_tilde = test_model(plot_x[0]).squeeze().detach().numpy()\n",
        "plot_x2 = plot_x.squeeze().numpy()\n",
        "ax.plot(plot_x2, plot_y_tilde, label = 'tune(disjoint)')\n",
        "ax.legend()\n",
        "fig.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tnc3Lr82wRtS",
        "colab_type": "code",
        "outputId": "57ad6c5a-59e4-4fae-ef32-833217e7d965",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "plot_y_tilde = pretrain(plot_x[0]).squeeze().detach().numpy()\n",
        "plot_x2 = plot_x.squeeze().numpy()\n",
        "ax.plot(plot_x2, plot_y_tilde, label = 'pretrain')\n",
        "ax.legend()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f18d86ac898>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U8SLgC1SAJnC",
        "colab_type": "text"
      },
      "source": [
        "執行底下的cell以顯示圖形，並重複執行更新meta weight與pretrained model的cell來比較多更新幾步後是否真的能看出meta learning比model pretraining有效"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jF5KQMHy6bt7",
        "colab_type": "code",
        "outputId": "87deaf94-8d3a-477c-ad44-1a06f30eeba4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 439
        }
      },
      "source": [
        "fig"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjsAAAGmCAYAAACaxkVbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCCGEEMKtSbIjhBBCCLf2/yH8m3pw+7aGAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 691.2x518.4 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    }
  ]
}