pytorch_workflow_model_exercise.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "uH0-Klx5r89r"
      },
      "outputs": [],
      "source": [
        "# import libraries\n",
        "import torch\n",
        "from torch import nn\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# check torch version\n",
        "torch.__version__"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "LvKOijS5sXLR",
        "outputId": "d171f9c0-6763-4d21-a151-a5d559f4b34b"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'2.2.1+cu121'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Setup agnostic code\n",
        "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
        "\n",
        "print(f\"Using device: {device}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OS3wE84Psaz4",
        "outputId": "609f5a6b-612f-4363-8f2a-befc0399b981"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Using device: cuda\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# data\n",
        "weight = 0.7\n",
        "bias = 0.3\n",
        "\n",
        "# create range of values\n",
        "start = 0\n",
        "end = 1\n",
        "step = 0.02\n",
        "\n",
        "# create X and y (features and labels)\n",
        "X = torch.arange(start, end, step).unsqueeze(dim=1)\n",
        "\n",
        "y = weight * X + bias"
      ],
      "metadata": {
        "id": "pSsyutwns5px"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Split data\n",
        "train_split = int(0.8 * len(X))\n",
        "X_train, y_train = X[:train_split], y[:train_split]\n",
        "X_test, y_test = X[train_split:], y[train_split:]"
      ],
      "metadata": {
        "id": "VgABMLuatZqp"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(X_train), len(y_train), len(X_test), len(y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-LlNFW_EtvSl",
        "outputId": "ad28b0bc-7abc-47d0-f8ad-12b4869bda08"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(40, 40, 10, 10)"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# create function for plot\n",
        "def plot_preds(train_data=X_train,\n",
        "               train_labels=y_train,\n",
        "               test_data=X_test,\n",
        "               test_labels=y_test,\n",
        "               predictions=None):\n",
        "  # plot train data\n",
        "  plt.scatter(train_data, train_labels, c=\"b\", s=4, label=\"Training Data\") # https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.scatter.html\n",
        "  # plot test data\n",
        "  plt.scatter(test_data, test_labels, c=\"g\", s=4, label=\"Test Data\")\n",
        "\n",
        "  # check if 'predictions' exist\n",
        "  if predictions is not None:\n",
        "    # plot the predictions\n",
        "    plt.scatter(test_data, predictions, c=\"r\", s=4, label=\"Predictions\")\n",
        "\n",
        "  # show the legend\n",
        "  plt.legend(prop={\"size\": 14})"
      ],
      "metadata": {
        "id": "-ScAPMcOt8tp"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_preds()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        },
        "id": "cCIhsdetwRog",
        "outputId": "28405990-9eed-4b35-bde9-4aad6421db5b"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create our model\n",
        "class LinearRegressionModelExercise(nn.Module):\n",
        "  def __init__(self):\n",
        "    super().__init__()\n",
        "    self.linear_layer = nn.Linear(in_features=1,\n",
        "                                   out_features=1)\n",
        "\n",
        "  def forward(self, x:torch.Tensor) -> torch.Tensor:\n",
        "    return self.linear_layer(x)"
      ],
      "metadata": {
        "id": "EyN_MkliwWWo"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# set the manual seed\n",
        "torch.manual_seed(42)\n",
        "\n",
        "# create instance of our model\n",
        "model = LinearRegressionModelExercise()\n",
        "\n",
        "model, model.state_dict()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3dIf5ykgcVd7",
        "outputId": "fc5cea4d-1e67-459d-bf98-9a1808a75858"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(LinearRegressionModelExercise(\n",
              "   (linear_layer): Linear(in_features=1, out_features=1, bias=True)\n",
              " ),\n",
              " OrderedDict([('linear_layer.weight', tensor([[0.7645]])),\n",
              "              ('linear_layer.bias', tensor([0.8300]))]))"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# check the model current device\n",
        "next(model.parameters()).device"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "v0AZtCUjcoED",
        "outputId": "7d59e561-2a90-45fe-bb0c-cbdb287e950b"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cpu')"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Set the model to use target device\n",
        "model.to(device)\n",
        "next(model.parameters()).device"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F6uQZWVjc4OE",
        "outputId": "2948b445-bfa6-43e6-8a13-af875eb3c5f7"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cuda', index=0)"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# for training our model we need 4 things on our hand: 1. loss function 2. optimizer 3. training loop and 4. testing loop\n",
        "# 1. loss function\n",
        "loss_fnc = nn.L1Loss()\n",
        "\n",
        "# 2. optimizer\n",
        "optimizer = torch.optim.SGD(model.parameters(),\n",
        "                            lr=0.01)\n"
      ],
      "metadata": {
        "id": "bhgJq9fedGpT"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 3. training loop\n",
        "torch.manual_seed(42)\n",
        "\n",
        "epochs=200\n",
        "\n",
        "# put the data to the target device (device agnosting code for data)\n",
        "X_train = X_train.to(device)\n",
        "y_train = y_train.to(device)\n",
        "X_test = X_test.to(device)\n",
        "y_test = y_test.to(device)\n",
        "\n",
        "for epoch in range(epochs):\n",
        "  model.train()\n",
        "\n",
        "  # 1. Forward pass\n",
        "  y_pred = model(X_train)\n",
        "\n",
        "  # 2. Calculate the loss\n",
        "  loss = loss_fnc(y_pred, y_train)\n",
        "\n",
        "  # 3. Optimizer zeor grad\n",
        "  optimizer.zero_grad()\n",
        "\n",
        "  # 4. backpropagation\n",
        "  loss.backward()\n",
        "\n",
        "  # 5. optimizer step\n",
        "  optimizer.step()\n",
        "\n",
        "\n",
        "  ### Testing\n",
        "  model.eval()\n",
        "\n",
        "  with torch.inference_mode():\n",
        "    test_pred = model(X_test)\n",
        "\n",
        "    # Calculate the loss\n",
        "    test_loss = loss_fnc(test_pred, y_test)\n",
        "\n",
        "  # print out the result\n",
        "  if epoch % 10 == 0:\n",
        "    print(f\"Epoch: {epoch} | Loss: {loss} | Test Loss | {test_loss}\")\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "figkQzKNeWyk",
        "outputId": "ed5dbe31-4d3b-4041-d054-283bba15b3a9"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 0 | Loss: 0.5551779866218567 | Test Loss | 0.5739762187004089\n",
            "Epoch: 10 | Loss: 0.439968079328537 | Test Loss | 0.4392664134502411\n",
            "Epoch: 20 | Loss: 0.3247582018375397 | Test Loss | 0.30455657839775085\n",
            "Epoch: 30 | Loss: 0.20954833924770355 | Test Loss | 0.16984669864177704\n",
            "Epoch: 40 | Loss: 0.09433845430612564 | Test Loss | 0.03513690456748009\n",
            "Epoch: 50 | Loss: 0.023886388167738914 | Test Loss | 0.04784907028079033\n",
            "Epoch: 60 | Loss: 0.019956795498728752 | Test Loss | 0.045803118497133255\n",
            "Epoch: 70 | Loss: 0.016517987474799156 | Test Loss | 0.037530567497015\n",
            "Epoch: 80 | Loss: 0.013089174404740334 | Test Loss | 0.02994490973651409\n",
            "Epoch: 90 | Loss: 0.009653178043663502 | Test Loss | 0.02167237363755703\n",
            "Epoch: 100 | Loss: 0.006215683650225401 | Test Loss | 0.014086711220443249\n",
            "Epoch: 110 | Loss: 0.00278724217787385 | Test Loss | 0.005814164876937866\n",
            "Epoch: 120 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n",
            "Epoch: 130 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n",
            "Epoch: 140 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n",
            "Epoch: 150 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n",
            "Epoch: 160 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n",
            "Epoch: 170 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n",
            "Epoch: 180 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n",
            "Epoch: 190 | Loss: 0.0012645035749301314 | Test Loss | 0.013801801018416882\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "weight, bias"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d2x1BOJCgI25",
        "outputId": "1287d2f0-0ed5-48eb-b173-1788ef2ffad5"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.7, 0.3)"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.state_dict()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vJkjXH-zg-zM",
        "outputId": "a52fadca-032c-4015-823e-1c29577c4c7a"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "OrderedDict([('linear_layer.weight', tensor([[0.6968]], device='cuda:0')),\n",
              "             ('linear_layer.bias', tensor([0.3025], device='cuda:0'))])"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# make predictions\n",
        "model.eval()\n",
        "with torch.inference_mode():\n",
        "  y_preds = model(X_test)\n",
        "y_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Mf7MJNFJhBJM",
        "outputId": "9f0c5f9f-21f2-49be-b63e-b381e9e505c0"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0.8600],\n",
              "        [0.8739],\n",
              "        [0.8878],\n",
              "        [0.9018],\n",
              "        [0.9157],\n",
              "        [0.9296],\n",
              "        [0.9436],\n",
              "        [0.9575],\n",
              "        [0.9714],\n",
              "        [0.9854]], device='cuda:0')"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_preds(predictions=y_preds.cpu())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        },
        "id": "13YSyx-vhxq1",
        "outputId": "7a7d70a4-b6a4-4748-9ac1-ebe708c46290"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 6.5 Saving and loading trained model\n",
        "\n",
        "There are three main methods you should about for saving and loading models in PyTorch.\n",
        "\n",
        "1. `torch.save()` - allows you to save a PyTorch object in [Python's pickle](https://docs.python.org/3/library/pickle.html) format\n",
        "2. `torch.load()` - allows you to load a PyTorch object\n",
        "3. `torch.nn.Module.load_state-dict()` - this allows to load a model's saved state dictionary"
      ],
      "metadata": {
        "id": "t09_0CyTmmuS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pathlib import Path\n",
        "\n",
        "# Create model directory\n",
        "MODEL_PATH = Path(\"models\")\n",
        "MODEL_PATH.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "# create a model save path\n",
        "MODEL_NAME = \"pytorch_workflow_model_exercise.pth\"\n",
        "MODEL_SAVE_PATH = MODEL_PATH / MODEL_NAME\n",
        "\n",
        "# save the model state dict\n",
        "print(f\"Saving model to: {MODEL_SAVE_PATH}\")\n",
        "torch.save(obj=model.state_dict(),\n",
        "           f= MODEL_SAVE_PATH)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e4dcauJuh4xE",
        "outputId": "e4cecb1d-285b-4b31-941f-76c171934db1"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Saving model to: models/pytorch_workflow_model_exercise.pth\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# loading the model\n",
        "loaded_model = LinearRegressionModelExercise()\n",
        "loaded_model.state_dict()"
      ],
      "metadata": {
        "id": "pEjXYnHenn-P",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e76dbb44-0d30-41c8-a39c-e7973d4c8c8a"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "OrderedDict([('linear_layer.weight', tensor([[0.7645]])),\n",
              "             ('linear_layer.bias', tensor([0.8300]))])"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# load the saved model state_dict\n",
        "loaded_model.load_state_dict(torch.load(f=MODEL_SAVE_PATH))\n",
        "\n",
        "# put the loaded model to device\n",
        "loaded_model.to(device)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FQyY4KxjGtdw",
        "outputId": "85a7dd26-ce4b-42d9-85d0-e1da92a5351e"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LinearRegressionModelExercise(\n",
              "  (linear_layer): Linear(in_features=1, out_features=1, bias=True)\n",
              ")"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# check parameters\n",
        "next(loaded_model.parameters()).device"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-zPi2d78HYrg",
        "outputId": "e757d1c7-2a28-4493-a08e-3f327bfdc132"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cuda', index=0)"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "loaded_model.state_dict()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QE8c4vTgHkyo",
        "outputId": "ced625d3-65a8-448b-9deb-60c1e7b46ed6"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "OrderedDict([('linear_layer.weight', tensor([[0.6968]], device='cuda:0')),\n",
              "             ('linear_layer.bias', tensor([0.3025], device='cuda:0'))])"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate the loaded model\n",
        "loaded_model.eval()\n",
        "\n",
        "# make predictions\n",
        "with torch.inference_mode():\n",
        "  y_pred_loaded_model = loaded_model(X_test)\n",
        "\n",
        "y_pred_loaded_model == y_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qKweAVRPHr0X",
        "outputId": "5e7eda7b-45e8-40ad-9fc9-092fd0ef1768"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[True],\n",
              "        [True],\n",
              "        [True],\n",
              "        [True],\n",
              "        [True],\n",
              "        [True],\n",
              "        [True],\n",
              "        [True],\n",
              "        [True],\n",
              "        [True]], device='cuda:0')"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# visualize\n",
        "plot_preds(predictions=y_pred_loaded_model.cpu())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        },
        "id": "P_HXwPozIPKv",
        "outputId": "c8a27f1c-1365-4e85-9843-ec0a887a76a3"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "vwUaD5DnIYL3"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}