How to take advantage of fastai with your pytorch training pipeline

pytorch2fastai.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "2021-02-14-Pytorchtofastai.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "6mDXFUG4sopG"
      },
      "source": [
        "#hide_input\n",
        "!pip install -Uqq fastai"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Qai8h9caUti_",
        "outputId": "c839117b-91cb-4295-bfe7-036ba5040cd4"
      },
      "source": [
        "import fastai\n",
        "print(fastai.__version__)"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2.5.2\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KGaGdX3DtLr6"
      },
      "source": [
        "import torch\n",
        "import torchvision\n",
        "import torchvision.transforms as transforms"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MP02xowOvrMs"
      },
      "source": [
        "transform = transforms.Compose(\n",
        "    [transforms.ToTensor(),\n",
        "     transforms.Normalize((0.5,0.5,0.5), (0.5,0.5,0.5))])"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FNHLOjJIv1uq",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "590e636d-d6c4-4d8d-caa0-3c5168e666ab"
      },
      "source": [
        "dset_train = torchvision.datasets.CIFAR10(root='./data', train=True,\n",
        "                                        download=True, transform=transform)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Files already downloaded and verified\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "La_ik28xv7tH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7e621445-8c5e-483e-a722-830a5afb1346"
      },
      "source": [
        "dset_test = torchvision.datasets.CIFAR10(root='./data', train=False,\n",
        "                                       download=True, transform=transform)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Files already downloaded and verified\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AH9tRI9mv9il"
      },
      "source": [
        "trainloader = torch.utils.data.DataLoader(dset_train, batch_size=4,\n",
        "                                          shuffle=True, num_workers=2)\n",
        "testloader = torch.utils.data.DataLoader(dset_test, batch_size=4,\n",
        "                                         shuffle=False, num_workers=2)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_2SI6qIvwwvN"
      },
      "source": [
        "import torch.nn as nn\n",
        "import torch.nn.functional as F"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LrXOQM5CxRYh"
      },
      "source": [
        "class Net(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(Net, self).__init__()\n",
        "        self.conv1 = nn.Conv2d(3, 6, 5)\n",
        "        self.pool = nn.MaxPool2d(2, 2)\n",
        "        self.conv2 = nn.Conv2d(6, 16, 5)\n",
        "        self.fc1 = nn.Linear(16 * 5 * 5, 120)\n",
        "        self.fc2 = nn.Linear(120, 84)\n",
        "        self.fc3 = nn.Linear(84, 10)\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.pool(F.relu(self.conv1(x)))\n",
        "        x = self.pool(F.relu(self.conv2(x)))\n",
        "        x = x.view(-1, 16 * 5 * 5)\n",
        "        x = F.relu(self.fc1(x))\n",
        "        x = F.relu(self.fc2(x))\n",
        "        x = self.fc3(x)\n",
        "        return x"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EC9Fc6A-xU96"
      },
      "source": [
        "net = Net()"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "54jt7ghrzOiP"
      },
      "source": [
        "criterion = nn.CrossEntropyLoss()"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XaBZiuQKza4o"
      },
      "source": [
        "from fastai.optimizer import OptimWrapper\n",
        "from torch import optim"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qcx1A4DTaW9a"
      },
      "source": [
        "def opt_func(params, **kwargs): return OptimWrapper(params, torch.optim.SGD, lr=0.001)"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ElZ85UkqxkBV"
      },
      "source": [
        "from fastai.data.core import DataLoaders"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iCJuAvO7x8U1"
      },
      "source": [
        "dls = DataLoaders(trainloader, testloader)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kKAndhNwraya"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_t9OdBCWyt0s"
      },
      "source": [
        "from fastai.learner import Learner"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8v6vrG9IMtT_"
      },
      "source": [
        "from fastai.callback.progress import ProgressCallback"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5f-StN4PyjlS"
      },
      "source": [
        "learn = Learner(dls, net, loss_func=criterion, opt_func=opt_func)"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 134
        },
        "id": "e7tOMYCv8G4j",
        "outputId": "378691be-f5d4-436f-a151-eb5336342d8b"
      },
      "source": [
        "learn.fit(1)"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: left;\">\n",
              "      <th>epoch</th>\n",
              "      <th>train_loss</th>\n",
              "      <th>valid_loss</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>2.111611</td>\n",
              "      <td>2.077898</td>\n",
              "      <td>01:04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/nn/functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at  /pytorch/c10/core/TensorImpl.h:1156.)\n",
            "  return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g7c6GjfnoW-C"
      },
      "source": [
        "torch.save(learn.model.state_dict(), './cifar_net.pth')"
      ],
      "execution_count": 20,
      "outputs": []
    }
  ]
}