task3/Protasov_Dmitry_ENAS.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# ENAS: Efficient Neural Architecture Search via Parameter Sharing\n",
        "\n",
        "Implement ENAS for CIFAR-10 dataset. Compare with random structure selection. Model: 3-layer network. Structure: binary mask for each parameter. Plots: model peformance, mask.\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "5NHKK2qpQ56K"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "$\\textbf{Training the shared parameters } \\omega \\textbf{ of the child models.}$ (= DynamicMLP)\n",
        "In this step, we fix the controller’s policy $\\pi(m;\\theta)$ and perform stochastic gradient descent (SGD) on $\\omega$ to minimize the expected loss function $\\mathbb{E}_{m\\sim\\pi(m;\\theta)} [L(m;\\omega)]$. Here, $L(m;\\omega)$ is the standard cross-entropy loss, computed on a minibatch of training data, with a model $m$ sampled from $\\pi(m;\\theta)$. The gradient is computed using the Monte Carlo estimate\n",
        "\n",
        "\\begin{equation}\n",
        "\\nabla_{\\omega} \\mathbb{E}_{m\\sim\\pi(m;\\theta)} [L(m;\\omega)] \\approx \\frac{1}{M} \\sum_{i=1}^{M} \\nabla_{\\omega} L(m_i;\\omega), \\quad (1)\n",
        "\\end{equation}\n",
        "\n",
        "where $m_i$'s are sampled from $\\pi(m;\\theta)$ (controller) as described above. Eqn 1 provides an unbiased estimate of the gradient $\\nabla_{\\omega} \\mathbb{E}_{m\\sim\\pi(m;\\theta)} [L(m;\\omega)]$. However, this estimate has a higher variance than the standard SGD gradient, where $m$ is fixed. Nevertheless – and this is perhaps surprising – we find that **$M = 1$ works just fine, i.e. we can update $\\omega$ using the gradient from any \\textit{single} model $m$ sampled from $\\pi(m;\\theta)$**. As mentioned, we train $\\omega$ during a entire pass through the training data.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "B_RXCrcFMBRr"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "$\\textbf{Training the controller parameters } \\theta$ (= LSTM). In this step, we fix $\\omega$ (DynamicMLP in my case) and update the policy parameters $\\theta$ (controller), aiming to maximize the expected reward $\\mathbb{E}_{m\\sim\\pi(m;\\theta)} [R(m,\\omega)]$. We employ the Adam optimizer, for which the gradient is computed using **REINFORCE algorithm** (Williams, 1992).\n",
        "\n",
        "The reward $R(m,\\omega)$ is computed on the validation set, rather than on the training set, to encourage ENAS to select models that generalize well rather than models that overfit the training set well. In our language model experiment, the reward function is $\\frac{C}{\\text{valid_ppl}}$, where the perplexity is computed on a minibatch of validation data. **In our image classification experiments, the reward function is the accuracy on a minibatch of validation images.**\n"
      ],
      "metadata": {
        "id": "9YfiZdMlMrm3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torchvision\n",
        "import torchvision.transforms as transforms\n",
        "\n",
        "transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])\n",
        "trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=transform)\n",
        "trainloader = torch.utils.data.DataLoader(trainset, batch_size=128, shuffle=True)\n",
        "\n",
        "testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=transform)\n",
        "testloader = torch.utils.data.DataLoader(testset, batch_size=128, shuffle=False)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DnME7JEJo6AX",
        "outputId": "e0d2f79d-3c66-40fd-a8da-52802bfeacb4"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "from torch.utils.data import DataLoader\n",
        "import torchvision.datasets as datasets\n",
        "import torchvision.transforms as transforms\n",
        "\n",
        "import torch.nn.functional as F\n",
        "\n",
        "\n",
        "class DynamicMLP(nn.Module):\n",
        "    def __init__(self, input_dim, max_layer1_neurons, max_layer2_neurons, output_dim):\n",
        "        super(DynamicMLP, self).__init__()\n",
        "        self.fc1 = nn.Linear(input_dim, max_layer1_neurons)\n",
        "        self.fc2 = nn.Linear(max_layer1_neurons, max_layer2_neurons)\n",
        "        self.fc3 = nn.Linear(max_layer2_neurons, output_dim)\n",
        "        self.mask_layer1 = None\n",
        "        self.mask_layer2 = None\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = F.relu(self.fc1(x.view(-1, input_dim)))\n",
        "        x = x * self.mask_layer1.expand_as(x)\n",
        "        x = F.relu(self.fc2(x))\n",
        "        x = x * self.mask_layer2.expand_as(x)\n",
        "        x = self.fc3(x)\n",
        "        return x\n",
        "\n",
        "    def apply_mask(self, mask_layer1, mask_layer2):\n",
        "        self.mask_layer1 = mask_layer1\n",
        "        self.mask_layer2 = mask_layer2"
      ],
      "metadata": {
        "id": "UrVMHvR32Dqj"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class Controller(nn.Module):\n",
        "    def __init__(self, max_layer1_neurons, max_layer2_neurons, hidden_dim, num_layers):\n",
        "        super(Controller, self).__init__()\n",
        "        self.num_layers = num_layers\n",
        "        self.hidden_dim = hidden_dim\n",
        "        self.lstm = nn.LSTM(input_size=3 * 32 * 32, hidden_size=hidden_dim, num_layers=num_layers)\n",
        "        self.fc = nn.Linear(hidden_dim, max_layer1_neurons + max_layer2_neurons)\n",
        "        self.log_probs = None\n",
        "\n",
        "    def forward(self, x):\n",
        "        batch_size = x.size(0)\n",
        "        h_0 = torch.zeros(self.num_layers, batch_size, self.hidden_dim)\n",
        "        c_0 = torch.zeros(self.num_layers, batch_size, self.hidden_dim)\n",
        "\n",
        "        out, (h_n, c_n) = self.lstm(x.view(batch_size, -1).unsqueeze(0), (h_0, c_0))\n",
        "        mask_logits = self.fc(h_n[-1])\n",
        "\n",
        "        mask_logits_layer1 = mask_logits[:, :max_layer1_neurons]\n",
        "        mask_logits_layer2 = mask_logits[:, max_layer1_neurons:max_layer1_neurons + max_layer2_neurons]\n",
        "        mask_layer1 = torch.sigmoid(mask_logits_layer1).mean(0) > 0.5\n",
        "        mask_layer2 = torch.sigmoid(mask_logits_layer2).mean(0) > 0.5\n",
        "\n",
        "        return mask_layer1.float(), mask_layer2.float(), mask_logits\n",
        "\n",
        "    def sample_masks(self):\n",
        "        # Не очень понял из статьи как ENAS сэмплирует архитектуру на инференсе\n",
        "        # решил подавать рандомный вход\n",
        "        # пробовал torch.zeros, отличия в экспериментах не заметил\n",
        "        x = torch.rand(1, 1, 3 * 32 * 32)\n",
        "        mask_layer1, mask_layer2, log_probs = self.forward(x)\n",
        "        self.log_probs = log_probs\n",
        "        return mask_layer1, mask_layer2"
      ],
      "metadata": {
        "id": "g8ZMwhPBi2bZ"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def train_shared_child_parameters(controller, mlp_model, trainloader, optimizer):\n",
        "    mlp_model.train()\n",
        "    for inputs, targets in trainloader:\n",
        "        optimizer.zero_grad()\n",
        "\n",
        "        inputs = inputs.view(-1, 3 * 32 * 32)\n",
        "        mask_layer1, mask_layer2, _ = controller.forward(inputs)\n",
        "\n",
        "        mlp_model.apply_mask(mask_layer1, mask_layer2)\n",
        "        outputs = mlp_model(inputs)\n",
        "        loss = F.cross_entropy(outputs, targets)\n",
        "\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "    return loss.item()"
      ],
      "metadata": {
        "id": "Xbf8gAMYpxui"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def evaluate(model, testloader):\n",
        "    model.eval()\n",
        "    accuracy = 0\n",
        "    total = 0\n",
        "\n",
        "    with torch.no_grad():\n",
        "        for x, y in testloader:\n",
        "            output = model(x)\n",
        "            _, predicted = torch.max(output, 1)\n",
        "            total += y.size(0)\n",
        "            accuracy += (predicted == y).sum().item()\n",
        "\n",
        "    return accuracy / total"
      ],
      "metadata": {
        "id": "rnKarhVDnMC5"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def train(controller, mlp_model, optimizer_mlp, optimizer_controller, trainloader, testloader, num_epochs, num_layers):\n",
        "    metrics = {'loss': [], 'reward': []}\n",
        "    for epoch in range(num_epochs):\n",
        "\n",
        "        # Train DynamicMLP\n",
        "        train_loss = train_shared_child_parameters(controller, mlp_model, trainloader, optimizer_mlp)\n",
        "\n",
        "        # Train Controller using REINFORCE\n",
        "        optimizer_controller.zero_grad()\n",
        "        mask_layer1, mask_layer2 = controller.sample_masks()\n",
        "        mlp_model.apply_mask(mask_layer1, mask_layer2)\n",
        "\n",
        "        reward = evaluate(mlp_model, testloader) # = Accuracy\n",
        "\n",
        "        policy_loss = [-log_prob * reward for log_prob in controller.log_probs]\n",
        "        policy_loss = torch.cat(policy_loss).sum()\n",
        "        policy_loss.backward()\n",
        "        optimizer_controller.step()\n",
        "\n",
        "        metrics['loss'].append(train_loss)\n",
        "        metrics['reward'].append(reward)\n",
        "        print(f'Epoch {epoch+1}, Total Loss: {train_loss:.6f}, Total Reward(Accuracy): {reward:.6f}')\n",
        "    return metrics"
      ],
      "metadata": {
        "id": "eWtEWkaJYVyh"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "input_dim = 3072  # Размер входных данных CIFAR-10\n",
        "hidden_dim = 100  # Размер скрытого слоя LSTM\n",
        "num_layers = 5  # Число слоев в LSTM\n",
        "output_dim = 10  # Количество классов в CIFAR-10\n",
        "max_layer1_neurons = 400\n",
        "max_layer2_neurons = 100\n",
        "\n",
        "\n",
        "controller = Controller(max_layer1_neurons, max_layer2_neurons, hidden_dim, num_layers)\n",
        "mlp_model = DynamicMLP(input_dim, max_layer1_neurons, max_layer2_neurons, output_dim)\n",
        "optimizer_mlp = optim.Adam(mlp_model.parameters(), lr=0.001)\n",
        "optimizer_controller = optim.Adam(controller.parameters(), lr=0.001)\n",
        "\n",
        "\n",
        "metrics = train(\n",
        "    controller, mlp_model, optimizer_mlp, optimizer_controller,\n",
        "    trainloader, testloader, num_epochs=20, num_layers=5\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R50gnwFpVx2o",
        "outputId": "a1cfb2c2-6f9f-4e9b-cac7-4e0fdadf8644"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1, Total Loss: 1.564681, Total Reward(Accuracy): 0.460600\n",
            "Epoch 2, Total Loss: 1.339337, Total Reward(Accuracy): 0.501500\n",
            "Epoch 3, Total Loss: 1.322967, Total Reward(Accuracy): 0.509300\n",
            "Epoch 4, Total Loss: 1.317593, Total Reward(Accuracy): 0.516700\n",
            "Epoch 5, Total Loss: 1.205146, Total Reward(Accuracy): 0.518200\n",
            "Epoch 6, Total Loss: 1.262240, Total Reward(Accuracy): 0.524900\n",
            "Epoch 7, Total Loss: 1.435195, Total Reward(Accuracy): 0.530600\n",
            "Epoch 8, Total Loss: 0.777352, Total Reward(Accuracy): 0.527000\n",
            "Epoch 9, Total Loss: 1.247403, Total Reward(Accuracy): 0.536100\n",
            "Epoch 10, Total Loss: 1.017258, Total Reward(Accuracy): 0.537800\n",
            "Epoch 11, Total Loss: 0.883572, Total Reward(Accuracy): 0.527000\n",
            "Epoch 12, Total Loss: 0.979680, Total Reward(Accuracy): 0.528100\n",
            "Epoch 13, Total Loss: 0.942416, Total Reward(Accuracy): 0.526900\n",
            "Epoch 14, Total Loss: 0.750811, Total Reward(Accuracy): 0.537900\n",
            "Epoch 15, Total Loss: 0.657746, Total Reward(Accuracy): 0.524000\n",
            "Epoch 16, Total Loss: 0.584060, Total Reward(Accuracy): 0.523100\n",
            "Epoch 17, Total Loss: 0.594000, Total Reward(Accuracy): 0.525800\n",
            "Epoch 18, Total Loss: 0.487284, Total Reward(Accuracy): 0.527200\n",
            "Epoch 19, Total Loss: 0.568221, Total Reward(Accuracy): 0.522500\n",
            "Epoch 20, Total Loss: 0.606020, Total Reward(Accuracy): 0.523200\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "fig, ax1 = plt.subplots(figsize=(12, 6))\n",
        "\n",
        "color = 'tab:red'\n",
        "ax1.set_xlabel('epoch')\n",
        "ax1.set_ylabel('loss', color=color)\n",
        "ax1.plot(metrics['loss'], color=color, label='loss')\n",
        "ax1.tick_params(axis='y', labelcolor=color)\n",
        "\n",
        "ax2 = ax1.twinx()\n",
        "color = 'tab:blue'\n",
        "ax2.set_ylabel('reward (accuracy)', color=color)\n",
        "ax2.plot(metrics['reward'], color=color, label='reward (accuracy)')\n",
        "ax2.tick_params(axis='y', labelcolor=color)\n",
        "\n",
        "fig.tight_layout()\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 606
        },
        "id": "XGD4xu-lznQB",
        "outputId": "9c518f43-52ba-4ed7-b25e-fa11526513e2"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x600 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tqdm import trange\n",
        "\n",
        "def RandomSelectionSearch(max_layer1_neurons, max_layer2_neurons):\n",
        "    mask_layer1 = torch.rand(max_layer1_neurons) > 0.5\n",
        "    mask_layer2 = torch.rand(max_layer2_neurons) > 0.5\n",
        "    return mask_layer1.float(), mask_layer2.float()\n",
        "\n",
        "rnd_mlp_model = DynamicMLP(input_dim, max_layer1_neurons, max_layer2_neurons, output_dim)\n",
        "num_iterations = 50\n",
        "\n",
        "best_reward = -float('inf')\n",
        "best_mask = RandomSelectionSearch(max_layer1_neurons, max_layer2_neurons)\n",
        "\n",
        "for i in trange(num_iterations):\n",
        "    mask_layer1, mask_layer2 = RandomSelectionSearch(max_layer1_neurons, max_layer2_neurons)\n",
        "\n",
        "    rnd_mlp_model.apply_mask(mask_layer1, mask_layer2)\n",
        "    reward = evaluate(rnd_mlp_model, testloader)\n",
        "\n",
        "    if reward > best_reward:\n",
        "        best_reward = reward\n",
        "        best_mask = mask_layer1, mask_layer2\n",
        "        print(f\"New Best Reward(Accuracy): {best_reward:.6f}\")\n",
        "\n",
        "rnd_mlp_model.apply_mask(*best_mask)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Px6gVGMvkprY",
        "outputId": "59ba55ad-ba34-45c1-ffd2-4d58d971c007"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "  2%|▏         | 1/50 [00:03<02:52,  3.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New Best Reward(Accuracy): 0.101300\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "  6%|▌         | 3/50 [00:11<02:56,  3.75s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New Best Reward(Accuracy): 0.102500\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            " 18%|█▊        | 9/50 [00:33<02:37,  3.84s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New Best Reward(Accuracy): 0.104300\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\r 20%|██        | 10/50 [00:37<02:30,  3.76s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New Best Reward(Accuracy): 0.111500\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\r 22%|██▏       | 11/50 [00:41<02:33,  3.94s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New Best Reward(Accuracy): 0.114100\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            " 42%|████▏     | 21/50 [01:29<02:11,  4.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New Best Reward(Accuracy): 0.121300\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            " 74%|███████▍  | 37/50 [02:33<00:49,  3.84s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New Best Reward(Accuracy): 0.123200\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 50/50 [03:24<00:00,  4.09s/it]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Видим, что RandomSelectionSearch не может получить модель с reward(accuracy) выше 0.13, в отличие от ENAS, который получает модель с accuracy примерно 0.5 уже после первой эпохи"
      ],
      "metadata": {
        "id": "-G2bjkX8y9ce"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "def visualize_masks(mask_layer1, mask_layer2):\n",
        "    mask_layer1_np = mask_layer1.cpu().numpy()\n",
        "    mask_layer2_np = mask_layer2.cpu().numpy()\n",
        "\n",
        "    # Визуализация маски для слоя 1\n",
        "    plt.figure(figsize=(20, 4))\n",
        "    plt.subplot(2, 1, 1)\n",
        "    plt.imshow(mask_layer1_np[np.newaxis, :], cmap='binary')\n",
        "    plt.title(\"Mask for Layer 1\")\n",
        "    plt.yticks([])\n",
        "\n",
        "    # Визуализация маски для слоя 2\n",
        "    plt.subplot(2, 1, 2)\n",
        "    plt.imshow(mask_layer2_np[np.newaxis, :], cmap='binary')\n",
        "    plt.title(\"Mask for Layer 2\")\n",
        "    plt.yticks([])\n",
        "\n",
        "    plt.show()\n",
        "\n",
        "visualize_masks(mlp_model.mask_layer1, mlp_model.mask_layer2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        },
        "id": "iR3aT2IpkkjT",
        "outputId": "a621c8e6-8c57-4beb-9c63-41fcf469b4dd"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x400 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# best_mask -- лучшие маски для RandomSelectionSearch\n",
        "visualize_masks(*best_mask)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        },
        "id": "2bQCeMppT71L",
        "outputId": "264484b5-d477-4506-c9f4-1020b40da498"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x400 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "- **ENAS Accuracy**: **0.523**.\n",
        "- **Random Selection Accuracy**: **0.123**.\n",
        "- Результаты ENAS не прям вау, возможно, стоит больше упереться в поиск оптимальных гиперпараметров\n",
        "\n",
        "**Conclusion**: ENAS shows promise with superior results to random selection but requires further hyperparameter optimization."
      ],
      "metadata": {
        "id": "59HWrTNTUzAO"
      }
    }
  ]
}