lesson5_sgd_mnist.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Copy of lesson5-sgd-mnist.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "zYZQTXtB1tkT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline\n",
        "from fastai.basics import *"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "342EjExx1tka",
        "colab_type": "text"
      },
      "source": [
        "## MNIST SGD"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9uQ23H7l1tkb",
        "colab_type": "text"
      },
      "source": [
        "Get the 'pickled' MNIST dataset from http://deeplearning.net/data/mnist/mnist.pkl.gz. We're going to treat it as a standard flat dataset with fully connected layers, rather than using a CNN."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dQrgggln4SId",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!mkdir /root/.fastai/data/mnist/"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8k0MjP3Q2Ags",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "4fc5f8f3-4d9d-48ec-edbf-398d7b60db61"
      },
      "source": [
        "!wget http://deeplearning.net/data/mnist/mnist.pkl.gz \n",
        "!mv mnist.pkl.gz /root/.fastai/data/mnist/\n",
        "!ls /root/.fastai/data/mnist/"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2019-06-16 15:34:51--  http://deeplearning.net/data/mnist/mnist.pkl.gz\n",
            "Resolving deeplearning.net (deeplearning.net)... 132.204.26.28\n",
            "Connecting to deeplearning.net (deeplearning.net)|132.204.26.28|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 16168813 (15M) [application/x-gzip]\n",
            "Saving to: ‘mnist.pkl.gz’\n",
            "\n",
            "mnist.pkl.gz        100%[===================>]  15.42M  3.43MB/s    in 6.7s    \n",
            "\n",
            "2019-06-16 15:34:58 (2.29 MB/s) - ‘mnist.pkl.gz’ saved [16168813/16168813]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZcqWF6vS1tkc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "path = Config().data_path()/'mnist'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cI2v9eey1tkg",
        "colab_type": "code",
        "outputId": "ac7aebcb-29d8-480d-df58-dc5dcbbcb171",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "path.ls()"
      ],
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[PosixPath('/root/.fastai/data/mnist/mnist.pkl.gz')]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NhLpxjVB1tkn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "with gzip.open(path/'mnist.pkl.gz', 'rb') as f:\n",
        "    ((x_train, y_train), (x_valid, y_valid), _) = pickle.load(f, encoding='latin-1')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qLLY5yT91tkq",
        "colab_type": "code",
        "outputId": "bec1ed20-b907-4acd-ab90-a73b27640959",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "plt.imshow(x_train[0].reshape((28,28)), cmap=\"gray\")\n",
        "x_train.shape"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(50000, 784)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 60
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hPM27XV51tkv",
        "colab_type": "code",
        "outputId": "71a0abb2-789a-4758-97ad-9fef0ed3902e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "x_train,y_train,x_valid,y_valid = map(torch.tensor, (x_train,y_train,x_valid,y_valid))\n",
        "n,c = x_train.shape\n",
        "x_train.shape, y_train.min(), y_train.max()"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(torch.Size([50000, 784]), tensor(0), tensor(9))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "O6q7PQyj1tkz",
        "colab_type": "text"
      },
      "source": [
        "In lesson2-sgd we did these things ourselves:\n",
        "\n",
        "```python\n",
        "x = torch.ones(n,2) \n",
        "def mse(y_hat, y): return ((y_hat-y)**2).mean()\n",
        "y_hat = x@a\n",
        "```\n",
        "\n",
        "Now instead we'll use PyTorch's functions to do it for us, and also to handle mini-batches (which we didn't do last time, since our dataset was so small)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f2DFK9xm1tk0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "bs=64\n",
        "train_ds = TensorDataset(x_train, y_train)\n",
        "valid_ds = TensorDataset(x_valid, y_valid)\n",
        "data = DataBunch.create(train_ds, valid_ds, bs=bs)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "04bi_ALI1tk3",
        "colab_type": "code",
        "outputId": "3cffa3a8-1911-4678-c376-6464dd16fd8c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "x,y = next(iter(data.train_dl))\n",
        "x.shape,y.shape"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(torch.Size([64, 784]), torch.Size([64]))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 63
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ieuQUmDN1tk8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Mnist_Logistic(nn.Module):\n",
        "    def __init__(self):\n",
        "        super().__init__()\n",
        "        self.lin = nn.Linear(784, 10, bias=True)\n",
        "\n",
        "    def forward(self, xb): return self.lin(xb)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W7qplWMO1tk_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = Mnist_Logistic().cuda()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Ri1vMcX1tlD",
        "colab_type": "code",
        "outputId": "32bd9088-ae72-4d88-dea7-4c22c310cbd4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "model"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Mnist_Logistic(\n",
              "  (lin): Linear(in_features=784, out_features=10, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cYILA3tN1tlG",
        "colab_type": "code",
        "outputId": "f873a91d-1703-4768-c244-206471949ce8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "model.lin"
      ],
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Linear(in_features=784, out_features=10, bias=True)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T36ytNca1tlK",
        "colab_type": "code",
        "outputId": "a7efa42b-f804-4d9b-c36b-efb5bd38f796",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "model(x).shape"
      ],
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([64, 10])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 68
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qRkzZMWo1tlO",
        "colab_type": "code",
        "outputId": "de4522c7-67d9-461d-d121-3b01985724fd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "[p.shape for p in model.parameters()]"
      ],
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[torch.Size([10, 784]), torch.Size([10])]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G781aNMg1tlS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "lr=2e-2"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EALnM3c41tlV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "loss_func = nn.CrossEntropyLoss()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CC-eZSPW1tlZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def update(x,y,lr):\n",
        "    wd = 1e-5\n",
        "    y_hat = model(x)\n",
        "    # weight decay\n",
        "    w2 = 0.\n",
        "    for p in model.parameters(): w2 += (p**2).sum()\n",
        "    # add to regular loss\n",
        "    loss = loss_func(y_hat, y) + w2*wd\n",
        "    loss.backward()\n",
        "    with torch.no_grad():\n",
        "        for p in model.parameters():\n",
        "            p.sub_(lr * p.grad)\n",
        "            p.grad.zero_()\n",
        "    return loss.item()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UVU2MQgr1tle",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "losses = [update(x,y,lr) for x,y in data.train_dl]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m_1KXhmK1tlh",
        "colab_type": "code",
        "outputId": "8643b648-eba6-43da-9e64-8d4adb52d3da",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.plot(losses);"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAD8CAYAAABw1c+bAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nB5u0J3PG2wKgAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UI6ODYnm1tlr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Mnist_NN(nn.Module):\n",
        "    def __init__(self):\n",
        "        super().__init__()\n",
        "        self.lin1 = nn.Linear(784, 50, bias=True)\n",
        "        self.lin2 = nn.Linear(50, 10, bias=True)\n",
        "\n",
        "    def forward(self, xb):\n",
        "        x = self.lin1(xb)\n",
        "        x = F.relu(x)\n",
        "        return self.lin2(x)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y-YPp2BH1tlu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = Mnist_NN().cuda()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0XDRe8_81tlw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "losses = [update(x,y,lr) for x,y in data.train_dl]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r6Rx6ttl1tl0",
        "colab_type": "code",
        "outputId": "843a24ad-a57d-45dc-9201-01ec6a795454",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.plot(losses);"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkVMAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SEdYumeb1tl3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = Mnist_NN().cuda()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5i_ufgmd1tl6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def update(x,y,lr):\n",
        "    opt = optim.Adam(model.parameters(), lr)\n",
        "    y_hat = model(x)\n",
        "    loss = loss_func(y_hat, y)\n",
        "    loss.backward()\n",
        "    opt.step()\n",
        "    opt.zero_grad()\n",
        "    return loss.item()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2d5MzpXs1tl8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "losses = [update(x,y,1e-3) for x,y in data.train_dl]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I0DSGfzb1tl-",
        "colab_type": "code",
        "outputId": "8a3241e3-734d-4bec-b2d0-932d7dbb89d5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.plot(losses);"
      ],
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4iEAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tqab59hG1tmD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "learn = Learner(data, Mnist_NN(), loss_func=loss_func, metrics=accuracy)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8uSyEcoN1tmH",
        "colab_type": "code",
        "outputId": "56349706-35df-42b1-c872-79407eb1eb31",
        "colab": {}
      },
      "source": [
        "%debug"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "> \u001b[0;32m/home/ubuntu/fastai/fastai/basic_data.py\u001b[0m(20)\u001b[0;36mDataLoader___getattr__\u001b[0;34m()\u001b[0m\n",
            "\u001b[0;32m     18 \u001b[0;31m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mutils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mDataLoader\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__init__\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mintercept_args\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m     19 \u001b[0;31m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m---> 20 \u001b[0;31m\u001b[0;32mdef\u001b[0m \u001b[0mDataLoader___getattr__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0mAny\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mgetattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdl\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m     21 \u001b[0;31m\u001b[0mDataLoader\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__getattr__\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mDataLoader___getattr__\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m     22 \u001b[0;31m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\n",
            "ipdb> u\n",
            "> \u001b[0;32m/home/ubuntu/fastai/fastai/basic_data.py\u001b[0m(38)\u001b[0;36m__getattr__\u001b[0;34m()\u001b[0m\n",
            "\u001b[0;32m     36 \u001b[0;31m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m     37 \u001b[0;31m    \u001b[0;32mdef\u001b[0m \u001b[0m__len__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0mint\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m---> 38 \u001b[0;31m    \u001b[0;32mdef\u001b[0m \u001b[0m__getattr__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0mAny\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mgetattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m     39 \u001b[0;31m    \u001b[0;32mdef\u001b[0m \u001b[0m__setstate__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mAny\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__dict__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\u001b[0;32m     40 \u001b[0;31m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0m\n",
            "ipdb> print(k)\n",
            "loss_func\n",
            "ipdb> q\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "henGl8Db1tmK",
        "colab_type": "code",
        "outputId": "a36db9e2-c4c3-4349-8912-41eff6b00a09",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        }
      },
      "source": [
        "learn.lr_find()\n",
        "learn.recorder.plot()"
      ],
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1CuvMehX1tmN",
        "colab_type": "code",
        "outputId": "e6a7beb2-43aa-46c4-af05-9214e3cd3633",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 80
        }
      },
      "source": [
        "learn.fit_one_cycle(1, 1e-2)"
      ],
      "execution_count": 85,
      "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>accuracy</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.142477</td>\n",
              "      <td>0.132384</td>\n",
              "      <td>0.962200</td>\n",
              "      <td>00:03</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KJUXr-4k1tmP",
        "colab_type": "code",
        "outputId": "4e5435fd-a042-4607-e739-2475c1e1b7bf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "source": [
        "learn.recorder.plot_lr(show_moms=True)"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuAAAAEKCAYAAABT6eBwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqQEXEREREYmg/wMcDgmBU7d63wAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 864x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iv5pc_Di1tmT",
        "colab_type": "code",
        "outputId": "b9f3d73a-9793-47a0-a381-af282eb6475d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "source": [
        "learn.recorder.plot_losses()"
      ],
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ACTMYLah1tmX",
        "colab_type": "text"
      },
      "source": [
        "## fin"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5TQy1j5Z1tmY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}