2_MLP.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rvsjBjiTaj0L"
      },
      "source": [
        "# 기초 MLP 실습\n",
        "\n",
        "Import torch"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "eIfKvf8Iagyc"
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "import torch.nn as nn"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gFhJYrRTcnNH"
      },
      "source": [
        "Device check(gpu)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OyCCshBmattT",
        "outputId": "ee0da713-8094-4e41-e682-ee8bf4a3c16f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "True\n",
            "cuda <class 'torch.device'>\n"
          ]
        }
      ],
      "source": [
        "# device check\n",
        "print(torch.cuda.is_available())\n",
        "\n",
        "if torch.cuda.is_available():\n",
        "  device = torch.device(\"cuda\")\n",
        "else:\n",
        "  device = torch.device(\"cpu\")\n",
        "\n",
        "print(device, type(device))\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qznrhl2tdXo3"
      },
      "source": [
        "## XOR problem\n",
        "\n",
        "![img](https://github.com/june-oh/2023_AI_Academy_ASR/blob/main/img/xor_table.jpg?raw=true)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "071EQk0KL-Vu"
      },
      "source": [
        "Dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "DengB8ASdJow"
      },
      "outputs": [],
      "source": [
        "# define X, Y\n",
        "X = [[0,0],[0,1],[1,0],[1,1]]\n",
        "Y = [[0],[1],[1],[0]]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "QC7F0kkAd0HW"
      },
      "outputs": [],
      "source": [
        "# define torch.tensor X, Y\n",
        "X=torch.tensor(X, dtype=torch.float)\n",
        "Y=torch.tensor(Y, dtype=torch.float)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oOLYHIasd3vH",
        "outputId": "5cb2f65c-f8a7-4537-8a99-259316125ec7"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(tensor([[0., 0.],\n",
              "         [0., 1.],\n",
              "         [1., 0.],\n",
              "         [1., 1.]]),\n",
              " torch.float32)"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "# check X and type of X\n",
        "X, X.dtype"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wzp2W9wwd4kE",
        "outputId": "255540e2-00b4-4554-f349-8465a734bf29"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(tensor([[0.],\n",
              "         [1.],\n",
              "         [1.],\n",
              "         [0.]]),\n",
              " torch.float32)"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ],
      "source": [
        "# check Y and type of Y\n",
        "Y, Y.dtype"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aPWP0Bd_eX59"
      },
      "source": [
        "## Perceptron\n",
        "\n",
        "<img src=\"https://github.com/june-oh/2023_AI_Academy_ASR/blob/main/img/perceptron.jpg?raw=true\" width=450>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "El1IYqkUex-R"
      },
      "source": [
        "### nn.sequential\n",
        "- 여러 neural network layer를 순서대로 적용하는 컨테이너\n",
        "\n",
        "### nn.linear\n",
        "- linear transform layer\n",
        "- arg\n",
        "  - in_features  : 입력 노드의 개수\n",
        "  - out_features : 출력 노드의 개수\n",
        "\n",
        "### nn.Sigmoid\n",
        "- sigmoid function\n",
        "- output = 0~1\n",
        "- used in binary classification problem"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "ckChgeltL-Vv"
      },
      "outputs": [],
      "source": [
        "??nn.Sequential"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "1EDyqpEgg4m-"
      },
      "outputs": [],
      "source": [
        "??nn.Linear"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "9xfRhe3gL-Vv"
      },
      "outputs": [],
      "source": [
        "??nn.Sigmoid"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GhOquD9RL-Vv"
      },
      "source": [
        "Define Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "wrlWPhBFd9Um"
      },
      "outputs": [],
      "source": [
        "#define basic model\n",
        "model = nn.Sequential(\n",
        "    nn.Linear(2, 1),\n",
        "    nn.Sigmoid()\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JX5V4dHUfTBh",
        "outputId": "0fc3ef94-4457-41e1-c70b-de91d5faa1dc"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential(\n",
              "  (0): Linear(in_features=2, out_features=1, bias=True)\n",
              "  (1): Sigmoid()\n",
              ")"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ],
      "source": [
        "# check model\n",
        "model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "it0MabxFfULs",
        "outputId": "8232071d-1216-4523-edc9-b7cf1b8a5539"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([0.6116], grad_fn=<SigmoidBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ],
      "source": [
        "# sample input\n",
        "sample_input = torch.tensor([0,0], dtype=torch.float)\n",
        "model(sample_input)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vrGwqHwFL-Vw"
      },
      "source": [
        "## Loss Function\n",
        "- 예측값과 실제 값 사이의 차이를 측정하는 함수\n",
        "- 이 손실 함수를 최소화하는 파라미터를 찾는 것이 딥러닝의 목적\n",
        "### nn.BCELoss\n",
        "- Binary Cross Entropy Loss\n",
        "- used in binary classification\n",
        "\n",
        "## Optimizer\n",
        "- 손실함수의 값을 최소화하는 모델의 파라미터를 찾는 알고리즘\n",
        "### optim.SGD\n",
        "- Stochastic Gradient Descent(SGD)\n",
        "    - Mini-Batch를 이용한 Gradient Descent\n",
        "### optim.Adam\n",
        "- Adaptive Moment Estimation(Adam)\n",
        "    - SGD + momentum + RMSprop\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "WLj-MSQLgUS9"
      },
      "outputs": [],
      "source": [
        "??torch.nn.BCELoss"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "7cCaF9-ggohW"
      },
      "outputs": [],
      "source": [
        "??torch.optim.SGD"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mrhh14ZIL-Vw"
      },
      "source": [
        "Define loss function and optimizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DgF7wLx6g_ml",
        "outputId": "501958d1-fb08-4a2b-f173-c345f3bb516d"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[Parameter containing:\n",
              " tensor([[ 0.0844, -0.3661]], requires_grad=True),\n",
              " Parameter containing:\n",
              " tensor([0.4541], requires_grad=True)]"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ],
      "source": [
        "# model parameter check\n",
        "list(model.parameters())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "nFQf_qJRf9kG"
      },
      "outputs": [],
      "source": [
        "# define loss_fn , optimizer\n",
        "loss_fn = torch.nn.BCELoss()\n",
        "optim = torch.optim.SGD(model.parameters(), lr=0.001)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HlnspUqWhQGA"
      },
      "source": [
        "### Train loop\n",
        "\n",
        "- epoch (전체데이터)\n",
        "  - batch (batch data)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49,
          "referenced_widgets": [
            "1ddaa7dfefb04f45aed979cddcd558d2",
            "65a762b1dfc04b46aa635e144f8217dc",
            "909b7890c5aa4228a41a3b94eb2b276b",
            "615a838c1be14677b5fb9157d88ca964",
            "f3861244dea0402f9cb236f83bc00d57",
            "c82435e87102494db89bb7a28072ee7f",
            "62b7b36ef1424fcea4d191875eca63ad",
            "75a2e2afc9174eef8d06ab2fc2a24f64",
            "d549200d5b0e4174aebc598bf895b83d",
            "238362b5a86e4122a66f30e3942e9e31",
            "e08c0cb39d01407d9832ad377ef95f51"
          ]
        },
        "id": "ARGvYXw-gk2y",
        "outputId": "f136bbf6-4ed0-4c58-9b50-35dad5643a84"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/10000 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "1ddaa7dfefb04f45aed979cddcd558d2"
            }
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Train loop\n",
        "from tqdm.auto import tqdm\n",
        "\n",
        "# epoch\n",
        "n_epoch= 10**4\n",
        "\n",
        "# model to device\n",
        "model.to(device)\n",
        "X = X.to(device)\n",
        "Y = Y.to(device)\n",
        "\n",
        "# loss tracking\n",
        "losses = []\n",
        "\n",
        "\n",
        "#training loop\n",
        "for epoch in tqdm(range(n_epoch)):\n",
        "  # gradient init\n",
        "  optim.zero_grad()\n",
        "\n",
        "  # prediction\n",
        "  pred = model(X)\n",
        "\n",
        "  # loss\n",
        "  loss = loss_fn(pred, Y)\n",
        "\n",
        "  # gradient backpropagation\n",
        "  loss.backward()\n",
        "\n",
        "  # loss tracking\n",
        "  if epoch % 10 ==0:\n",
        "    losses.append(loss.item())\n",
        "\n",
        "  # model parameter update\n",
        "  optim.step()\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 448
        },
        "id": "uJtOJyrchzNQ",
        "outputId": "346ded0c-966a-42cd-9c2d-078d5cf3029b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.6958434581756592\n"
          ]
        }
      ],
      "source": [
        "# check loss\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "plt.plot(losses)\n",
        "plt.show()\n",
        "\n",
        "print(losses[-1])\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aedNFGf0h1N2",
        "outputId": "ca3ae789-733e-44d6-d234-1743d8d20c02"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0.5490],\n",
              "        [0.4787],\n",
              "        [0.5391],\n",
              "        [0.4687]], device='cuda:0', grad_fn=<SigmoidBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "# Inference\n",
        "model(X)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oT2mm2jKohYj"
      },
      "source": [
        "## define MLP model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "zeS9bAZ5jjRV"
      },
      "outputs": [],
      "source": [
        "# define MLP model\n",
        "mlp_model = nn.Sequential(\n",
        "    nn.Linear(2,2,bias=True),\n",
        "    nn.Sigmoid(),\n",
        "    nn.Linear(2,1,bias=True),\n",
        "    nn.Sigmoid()\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "mlp_model"
      ],
      "metadata": {
        "id": "xo0lZ4AJWhVL",
        "outputId": "c9e199d2-38ab-4fdb-8e5f-856845f41a03",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential(\n",
              "  (0): Linear(in_features=2, out_features=2, bias=True)\n",
              "  (1): Sigmoid()\n",
              "  (2): Linear(in_features=2, out_features=1, bias=True)\n",
              "  (3): Sigmoid()\n",
              ")"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D3ymtAj8L-Vw"
      },
      "source": [
        "모델을 학습할때 lr을 충분히 크게(e.g., lr = 0.1) 줘야 학습이 용이"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "id": "ErI0Ax3WL-Vx"
      },
      "outputs": [],
      "source": [
        "# define loss function and optimizer\n",
        "loss_fn = torch.nn.BCELoss()\n",
        "optim = torch.optim.Adam(mlp_model.parameters(), lr=0.01)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49,
          "referenced_widgets": [
            "5637170642c7446a894becf1b0c90580",
            "ce4a77d600844ba784b701466b23332c",
            "1e3d6fb21ed74204b854517e2eede20f",
            "d65a530322e040c6b5eadd76d76872ce",
            "7d45bcbcfc154306bf5c0e1b8af7d210",
            "68c05ee05e34466fbcce25c8a76df66b",
            "ad85569ee37645fdac503dcdfb41011a",
            "00f1cd3697f142b5a126c85b53f371f7",
            "409300b5193a4133a1723dd749a09d4f",
            "970e3843210947d0a2762bcdaba13161",
            "00fe101f3ec449d4a87a1aa298eb6cdc"
          ]
        },
        "id": "92zA-wxLpePz",
        "outputId": "a5246330-febf-4c73-ab4d-4740bdf5bc11"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/10000 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "5637170642c7446a894becf1b0c90580"
            }
          },
          "metadata": {}
        }
      ],
      "source": [
        "\n",
        "\n",
        "# Train loop\n",
        "# epoch\n",
        "n_epoch= 10**4\n",
        "\n",
        "# model to device\n",
        "mlp_model.to(device)\n",
        "X = X.to(device)\n",
        "Y = Y.to(device)\n",
        "\n",
        "# loss tracking\n",
        "losses = []\n",
        "\n",
        "\n",
        "pbar = tqdm(range(n_epoch))\n",
        "for epoch in pbar:\n",
        "  # gradient init\n",
        "  optim.zero_grad()\n",
        "\n",
        "  # prediction\n",
        "  pred = mlp_model(X)\n",
        "\n",
        "  # loss\n",
        "  loss = loss_fn(pred, Y)\n",
        "\n",
        "  # gradient backpropagation\n",
        "  loss.backward()\n",
        "\n",
        "  pbar.desc = f\"loss : {loss.item()}\"\n",
        "\n",
        "  # loss tracking\n",
        "  if epoch % 10 ==0:\n",
        "    losses.append(loss.item())\n",
        "\n",
        "  # model parameter update\n",
        "  optim.step()\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 448
        },
        "id": "uE4NO4tGp82i",
        "outputId": "d97623b1-3178-4dac-a1e4-2d48e68a72c3"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "8.157946285791695e-05\n"
          ]
        }
      ],
      "source": [
        "# check loss\n",
        "plt.plot(losses)\n",
        "plt.show()\n",
        "\n",
        "print(losses[-1])\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DS2fVrnlqAYa",
        "outputId": "228c9145-828d-4b3e-87de-1fcee7d924cd"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[    0.00007816125435056165],\n",
              "        [    0.99992001056671142578],\n",
              "        [    0.99991178512573242188],\n",
              "        [    0.00007831882248865440]], device='cuda:0',\n",
              "       grad_fn=<SigmoidBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 31
        }
      ],
      "source": [
        "# Inference\n",
        "torch.set_printoptions(precision=20,sci_mode=False)\n",
        "mlp_model(X)"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "CqAhck6gX3kW"
      },
      "execution_count": 25,
      "outputs": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.12"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "1ddaa7dfefb04f45aed979cddcd558d2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_65a762b1dfc04b46aa635e144f8217dc",
              "IPY_MODEL_909b7890c5aa4228a41a3b94eb2b276b",
              "IPY_MODEL_615a838c1be14677b5fb9157d88ca964"
            ],
            "layout": "IPY_MODEL_f3861244dea0402f9cb236f83bc00d57"
          }
        },
        "65a762b1dfc04b46aa635e144f8217dc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c82435e87102494db89bb7a28072ee7f",
            "placeholder": "​",
            "style": "IPY_MODEL_62b7b36ef1424fcea4d191875eca63ad",
            "value": "100%"
          }
        },
        "909b7890c5aa4228a41a3b94eb2b276b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_75a2e2afc9174eef8d06ab2fc2a24f64",
            "max": 10000,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_d549200d5b0e4174aebc598bf895b83d",
            "value": 10000
          }
        },
        "615a838c1be14677b5fb9157d88ca964": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_238362b5a86e4122a66f30e3942e9e31",
            "placeholder": "​",
            "style": "IPY_MODEL_e08c0cb39d01407d9832ad377ef95f51",
            "value": " 10000/10000 [00:10&lt;00:00, 1545.33it/s]"
          }
        },
        "f3861244dea0402f9cb236f83bc00d57": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c82435e87102494db89bb7a28072ee7f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "62b7b36ef1424fcea4d191875eca63ad": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "75a2e2afc9174eef8d06ab2fc2a24f64": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "d549200d5b0e4174aebc598bf895b83d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "238362b5a86e4122a66f30e3942e9e31": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "e08c0cb39d01407d9832ad377ef95f51": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "5637170642c7446a894becf1b0c90580": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_ce4a77d600844ba784b701466b23332c",
              "IPY_MODEL_1e3d6fb21ed74204b854517e2eede20f",
              "IPY_MODEL_d65a530322e040c6b5eadd76d76872ce"
            ],
            "layout": "IPY_MODEL_7d45bcbcfc154306bf5c0e1b8af7d210"
          }
        },
        "ce4a77d600844ba784b701466b23332c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_68c05ee05e34466fbcce25c8a76df66b",
            "placeholder": "​",
            "style": "IPY_MODEL_ad85569ee37645fdac503dcdfb41011a",
            "value": "loss : 8.119398989947513e-05: 100%"
          }
        },
        "1e3d6fb21ed74204b854517e2eede20f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_00f1cd3697f142b5a126c85b53f371f7",
            "max": 10000,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_409300b5193a4133a1723dd749a09d4f",
            "value": 10000
          }
        },
        "d65a530322e040c6b5eadd76d76872ce": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_970e3843210947d0a2762bcdaba13161",
            "placeholder": "​",
            "style": "IPY_MODEL_00fe101f3ec449d4a87a1aa298eb6cdc",
            "value": " 10000/10000 [00:12&lt;00:00, 915.51it/s]"
          }
        },
        "7d45bcbcfc154306bf5c0e1b8af7d210": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "68c05ee05e34466fbcce25c8a76df66b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ad85569ee37645fdac503dcdfb41011a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "00f1cd3697f142b5a126c85b53f371f7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "409300b5193a4133a1723dd749a09d4f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "970e3843210947d0a2762bcdaba13161": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "00fe101f3ec449d4a87a1aa298eb6cdc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}