Created using Colaboratory

sjnaj · May 9, 2022 · 5adc5fc · 5adc5fc
1 parent 92885ef
commit 5adc5fc
Showing 1 changed file with 301 additions and 0 deletions.
diff --git a/Basic/layersAndBlocks.ipynb b/Basic/layersAndBlocks.ipynb
@@ -0,0 +1,301 @@
+{
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
+    "colab": {
+      "name": "Untitled3.ipynb",
+      "provenance": [],
+      "collapsed_sections": [],
+      "authorship_tag": "ABX9TyMZFdaZxqHzr/+baEamEDGw",
+      "include_colab_link": true
+    },
+    "kernelspec": {
+      "name": "python3",
+      "display_name": "Python 3"
+    },
+    "language_info": {
+      "name": "python"
+    }
+  },
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "view-in-github",
+        "colab_type": "text"
+      },
+      "source": [
+        "<a href=\"https://colab.research.google.com/github/sjnaj/DeepLearning/blob/master/Basic/layersAndBlocks.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 2,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "zlmpgLiM8BNI",
+        "outputId": "4ebc37ad-592d-44dd-9710-07e7646cfeb4"
+      },
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "tensor([[-0.0708,  0.1940, -0.1375, -0.1126, -0.0094, -0.2533, -0.0626,  0.0696,\n",
+              "         -0.2112,  0.1389],\n",
+              "        [-0.0099,  0.0608, -0.0250, -0.0163,  0.0969, -0.3051, -0.1870, -0.1103,\n",
+              "         -0.1439,  0.1661]], grad_fn=<AddmmBackward0>)"
+            ]
+          },
+          "metadata": {},
+          "execution_count": 2
+        }
+      ],
+      "source": [
+        "import torch\n",
+        "from torch import nn\n",
+        "from torch.nn import functional as F\n",
+        "\n",
+        "net = nn.Sequential(nn.Linear(20, 256), nn.ReLU(), nn.Linear(256, 10))\n",
+        "\n",
+        "X = torch.rand(2, 20)\n",
+        "net(X)\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "自定义块"
+      ],
+      "metadata": {
+        "id": "2kP9S3ChUmY-"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "class MLP(nn.Module):\n",
+        "    # 用模型参数声明层。这里，我们声明两个全连接的层\n",
+        "    def __init__(self):\n",
+        "        # 调用MLP的父类Module的构造函数来执行必要的初始化。\n",
+        "        # 这样，在类实例化时也可以指定其他函数参数，例如模型参数params（稍后将介绍）\n",
+        "        super().__init__()\n",
+        "        self.hidden = nn.Linear(20, 256)  # 隐藏层\n",
+        "        self.out = nn.Linear(256, 10)  # 输出层\n",
+        "\n",
+        "    # 定义模型的前向传播，即如何根据输入X返回所需的模型输出\n",
+        "    def forward(self, X):#在nn.Module中映射到__self__上\n",
+        "        # 注意，这里我们使用ReLU的函数版本，其在nn.functional模块中定义。\n",
+        "        return self.out(F.relu(self.hidden(X)))"
+      ],
+      "metadata": {
+        "id": "c0hx3TRQR0-q"
+      },
+      "execution_count": 3,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "net=MLP()\n",
+        "net(X)"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "0sDm8w2OTu_s",
+        "outputId": "97d9ecbd-8951-4b0e-e42b-72293e46c0c6"
+      },
+      "execution_count": 4,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "tensor([[-0.0938, -0.3164,  0.1827,  0.0036,  0.0163, -0.1917, -0.2530,  0.0999,\n",
+              "          0.0322, -0.0096],\n",
+              "        [-0.0999, -0.1062,  0.1479, -0.0119,  0.1503,  0.0283, -0.1485,  0.0617,\n",
+              "          0.0884, -0.1171]], grad_fn=<AddmmBackward0>)"
+            ]
+          },
+          "metadata": {},
+          "execution_count": 4
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "自定义顺序块"
+      ],
+      "metadata": {
+        "id": "KMYB6ntWUpPk"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "class MySequential(nn.Module):\n",
+        "    def __init__(self, *args):\n",
+        "        super().__init__()\n",
+        "        for idx, module in enumerate(args):\n",
+        "            # 这里，module是Module子类的一个实例。我们把它保存在'Module'类的成员\n",
+        "            # 变量_modules中。module的类型是OrderedDict\n",
+        "            self._modules[str(idx)] = module\n",
+        "\n",
+        "    def forward(self, X):\n",
+        "        # OrderedDict保证了按照成员添加的顺序遍历它们\n",
+        "        for block in self._modules.values():\n",
+        "            X = block(X)\n",
+        "        return X\n"
+      ],
+      "metadata": {
+        "id": "6DwsPggET1en"
+      },
+      "execution_count": 6,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "net = MySequential(nn.Linear(20, 256), nn.ReLU(), nn.Linear(256, 10))\n",
+        "net(X)"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "Acs23gnNUF74",
+        "outputId": "57737a2a-3801-4635-c919-f684f97e7455"
+      },
+      "execution_count": 7,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "tensor([[-0.3066, -0.0166, -0.0799, -0.0569,  0.2381,  0.0321,  0.0371,  0.0777,\n",
+              "          0.0003, -0.0135],\n",
+              "        [-0.0912, -0.0402, -0.0250,  0.0942,  0.1888, -0.0461,  0.0055,  0.1226,\n",
+              "          0.1062, -0.1039]], grad_fn=<AddmmBackward0>)"
+            ]
+          },
+          "metadata": {},
+          "execution_count": 7
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "灵活的forward"
+      ],
+      "metadata": {
+        "id": "EuvT1RU6Uvwk"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "class FixedHiddenMLP(nn.Module):\n",
+        "    def __init__(self):\n",
+        "        super().__init__()\n",
+        "        # 不计算梯度的随机权重参数。因此其在训练期间保持不变\n",
+        "        self.rand_weight = torch.rand((20, 20), requires_grad=False)#False当作标量，否则当作变量\n",
+        "        self.linear = nn.Linear(20, 20)\n",
+        "\n",
+        "    def forward(self, X):\n",
+        "        X = self.linear(X)\n",
+        "        # 使用创建的常量参数以及relu和mm函数\n",
+        "        X = F.relu(torch.mm(X, self.rand_weight) + 1)\n",
+        "        # 复用全连接层。这相当于两个全连接层共享参数\n",
+        "        X = self.linear(X)\n",
+        "        # 控制流\n",
+        "        while X.abs().sum() > 1:\n",
+        "            X /= 2\n",
+        "        return X.sum()"
+      ],
+      "metadata": {
+        "id": "ZsVZO8m1UMK8"
+      },
+      "execution_count": 9,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "net = FixedHiddenMLP()\n",
+        "net(X)"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "wJlXhHm4UdD6",
+        "outputId": "ac65e806-e10c-4af1-85ad-a6d3d9f79030"
+      },
+      "execution_count": 10,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "tensor(-0.2742, grad_fn=<SumBackward0>)"
+            ]
+          },
+          "metadata": {},
+          "execution_count": 10
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "混搭各种块"
+      ],
+      "metadata": {
+        "id": "1-zmHy59U23S"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "class NestMLP(nn.Module):\n",
+        "    def __init__(self):\n",
+        "        super().__init__()\n",
+        "        self.net = nn.Sequential(nn.Linear(20, 64), nn.ReLU(),\n",
+        "                                 nn.Linear(64, 32), nn.ReLU())\n",
+        "        self.linear = nn.Linear(32, 16)\n",
+        "\n",
+        "    def forward(self, X):\n",
+        "        return self.linear(self.net(X))\n",
+        "\n",
+        "chimera = nn.Sequential(NestMLP(), nn.Linear(16, 20), FixedHiddenMLP())\n",
+        "chimera(X)\n"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "Pa7Jgy7fUg9R",
+        "outputId": "eae4755f-9193-469b-f11c-84755f0f24bc"
+      },
+      "execution_count": 11,
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "tensor(-0.2437, grad_fn=<SumBackward0>)"
+            ]
+          },
+          "metadata": {},
+          "execution_count": 11
+        }
+      ]
+    }
+  ]
+}