FedAvg_MNIST_ANN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IsdzjGPzoAjk",
        "outputId": "ce96b388-7862-4172-eb75-6e76a27dacb0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz to data/MNIST/raw/train-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 9912422/9912422 [00:00<00:00, 136236532.21it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/MNIST/raw/train-images-idx3-ubyte.gz to data/MNIST/raw\n",
            "\n",
            "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz to data/MNIST/raw/train-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 28881/28881 [00:00<00:00, 16722210.63it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/MNIST/raw/train-labels-idx1-ubyte.gz to data/MNIST/raw\n",
            "\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz to data/MNIST/raw/t10k-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 1648877/1648877 [00:00<00:00, 32444754.37it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/MNIST/raw/t10k-images-idx3-ubyte.gz to data/MNIST/raw\n",
            "\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz to data/MNIST/raw/t10k-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 4542/4542 [00:00<00:00, 4455221.88it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/MNIST/raw/t10k-labels-idx1-ubyte.gz to data/MNIST/raw\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train Images Shape: torch.Size([60000, 28, 28])\n",
            "Train Labels Shape: torch.Size([60000])\n",
            "Test Images Shape: torch.Size([10000, 28, 28])\n",
            "Test Labels Shape: torch.Size([10000])\n"
          ]
        }
      ],
      "source": [
        "import torchvision.datasets as datasets\n",
        "\n",
        "# Load MNIST dataset\n",
        "train_dataset = datasets.MNIST(root='data/', train=True, download=True)\n",
        "test_dataset = datasets.MNIST(root='data/', train=False, download=True)\n",
        "\n",
        "# Split dataset into test and train\n",
        "train_images = train_dataset.data\n",
        "train_labels = train_dataset.targets\n",
        "test_images = test_dataset.data\n",
        "test_labels = test_dataset.targets\n",
        "\n",
        "# Print shape of the datasets\n",
        "print(\"Train Images Shape:\", train_images.shape)\n",
        "print(\"Train Labels Shape:\", train_labels.shape)\n",
        "print(\"Test Images Shape:\", test_images.shape)\n",
        "print(\"Test Labels Shape:\", test_labels.shape)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install torch"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nePKZLqitVvn",
        "outputId": "655f654f-1ba5-4a85-8724-eed8172096c5"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: torch in /usr/local/lib/python3.10/dist-packages (2.0.1+cu118)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch) (3.12.2)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from torch) (4.7.1)\n",
            "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch) (1.11.1)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch) (3.1)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch) (3.1.2)\n",
            "Requirement already satisfied: triton==2.0.0 in /usr/local/lib/python3.10/dist-packages (from torch) (2.0.0)\n",
            "Requirement already satisfied: cmake in /usr/local/lib/python3.10/dist-packages (from triton==2.0.0->torch) (3.25.2)\n",
            "Requirement already satisfied: lit in /usr/local/lib/python3.10/dist-packages (from triton==2.0.0->torch) (16.0.6)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch) (2.1.3)\n",
            "Requirement already satisfied: mpmath>=0.19 in /usr/local/lib/python3.10/dist-packages (from sympy->torch) (1.3.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install tensorflow\n",
        "import tensorflow as tf"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H7XgioOJtlZS",
        "outputId": "51c30220-f6f1-4b93-b454-1fb016fd5315"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: tensorflow in /usr/local/lib/python3.10/dist-packages (2.12.0)\n",
            "Requirement already satisfied: absl-py>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (1.4.0)\n",
            "Requirement already satisfied: astunparse>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (1.6.3)\n",
            "Requirement already satisfied: flatbuffers>=2.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (23.5.26)\n",
            "Requirement already satisfied: gast<=0.4.0,>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (0.4.0)\n",
            "Requirement already satisfied: google-pasta>=0.1.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (0.2.0)\n",
            "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (1.56.0)\n",
            "Requirement already satisfied: h5py>=2.9.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (3.8.0)\n",
            "Requirement already satisfied: jax>=0.3.15 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (0.4.13)\n",
            "Requirement already satisfied: keras<2.13,>=2.12.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (2.12.0)\n",
            "Requirement already satisfied: libclang>=13.0.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (16.0.0)\n",
            "Requirement already satisfied: numpy<1.24,>=1.22 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (1.22.4)\n",
            "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (3.3.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from tensorflow) (23.1)\n",
            "Requirement already satisfied: protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<5.0.0dev,>=3.20.3 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (3.20.3)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from tensorflow) (67.7.2)\n",
            "Requirement already satisfied: six>=1.12.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (1.16.0)\n",
            "Requirement already satisfied: tensorboard<2.13,>=2.12 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (2.12.3)\n",
            "Requirement already satisfied: tensorflow-estimator<2.13,>=2.12.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (2.12.0)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (2.3.0)\n",
            "Requirement already satisfied: typing-extensions>=3.6.6 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (4.7.1)\n",
            "Requirement already satisfied: wrapt<1.15,>=1.11.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (1.14.1)\n",
            "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow) (0.32.0)\n",
            "Requirement already satisfied: wheel<1.0,>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from astunparse>=1.6.0->tensorflow) (0.40.0)\n",
            "Requirement already satisfied: ml-dtypes>=0.1.0 in /usr/local/lib/python3.10/dist-packages (from jax>=0.3.15->tensorflow) (0.2.0)\n",
            "Requirement already satisfied: scipy>=1.7 in /usr/local/lib/python3.10/dist-packages (from jax>=0.3.15->tensorflow) (1.10.1)\n",
            "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.13,>=2.12->tensorflow) (2.17.3)\n",
            "Requirement already satisfied: google-auth-oauthlib<1.1,>=0.5 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.13,>=2.12->tensorflow) (1.0.0)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.13,>=2.12->tensorflow) (3.4.3)\n",
            "Requirement already satisfied: requests<3,>=2.21.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.13,>=2.12->tensorflow) (2.27.1)\n",
            "Requirement already satisfied: tensorboard-data-server<0.8.0,>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.13,>=2.12->tensorflow) (0.7.1)\n",
            "Requirement already satisfied: werkzeug>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.13,>=2.12->tensorflow) (2.3.6)\n",
            "Requirement already satisfied: cachetools<6.0,>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (5.3.1)\n",
            "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (0.3.0)\n",
            "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (4.9)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from google-auth-oauthlib<1.1,>=0.5->tensorboard<2.13,>=2.12->tensorflow) (1.3.1)\n",
            "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (1.26.16)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (2023.5.7)\n",
            "Requirement already satisfied: charset-normalizer~=2.0.0 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (2.0.12)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (3.4)\n",
            "Requirement already satisfied: MarkupSafe>=2.1.1 in /usr/local/lib/python3.10/dist-packages (from werkzeug>=1.0.1->tensorboard<2.13,>=2.12->tensorflow) (2.1.3)\n",
            "Requirement already satisfied: pyasn1<0.6.0,>=0.4.6 in /usr/local/lib/python3.10/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (0.5.0)\n",
            "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.10/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<1.1,>=0.5->tensorboard<2.13,>=2.12->tensorflow) (3.2.2)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_images = tf.convert_to_tensor(train_images.numpy(), dtype=tf.float32)\n",
        "train_labels = tf.convert_to_tensor(train_labels.numpy(), dtype=tf.int32)"
      ],
      "metadata": {
        "id": "w4aTHmTcAIlr"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Equal division of data"
      ],
      "metadata": {
        "id": "P--DgEcm8dNw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Split it into groups of 4 because 60000 divvides evenly among 4 groups\n",
        "split_images = tf.split(train_images, 4, axis=0)\n",
        "split_labels = tf.split(train_labels, 4, axis=0)"
      ],
      "metadata": {
        "id": "mfCj_eSgpv-t"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import torchvision.transforms as transforms\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "oTa6grBauOiT"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(len(split_images)):\n",
        "    print(f\"Group {i+1} - Images: {split_images[i].shape}, Labels: {split_labels[i].shape}\")\n",
        "    fig, axs = plt.subplots(1, 4, figsize=(12, 3))\n",
        "    for j in range(4):\n",
        "        img = split_images[i][j]\n",
        "        label = split_labels[i][j]\n",
        "        axs[j].imshow(img, cmap='gray')\n",
        "        axs[j].set_title(f\"Label: {label}\")\n",
        "        axs[j].axis('off')\n",
        "    plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "bL8gWx4Strdr",
        "outputId": "68d6f600-e35a-4d96-de00-30c6f32294d2"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Group 1 - Images: (15000, 28, 28), Labels: (15000,)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x300 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Group 2 - Images: (15000, 28, 28), Labels: (15000,)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x300 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Group 3 - Images: (15000, 28, 28), Labels: (15000,)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x300 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Group 4 - Images: (15000, 28, 28), Labels: (15000,)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x300 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# our model architecture\n",
        "model = tf.keras.Sequential([\n",
        "    tf.keras.layers.Flatten(input_shape=(28, 28)),\n",
        "    tf.keras.layers.Dense(128, activation='relu'),\n",
        "    tf.keras.layers.Dense(10, activation='softmax')\n",
        "])"
      ],
      "metadata": {
        "id": "TSUQlVztwmBO"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np"
      ],
      "metadata": {
        "id": "SlXk_dD4w77l"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Random Seeding"
      ],
      "metadata": {
        "id": "Vmc49ZE37xkQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "\n",
        "trained_models = []\n",
        "group_weights = []  # List to store the weights of each group\n",
        "\n",
        "for i in range(len(split_images)):\n",
        "    print(f\"Training model for Group {i+1}\")\n",
        "    seed = np.random.randint(0, 1000)\n",
        "    tf.random.set_seed(seed)\n",
        "    model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "    model.fit(split_images[i], split_labels[i], epochs=5, batch_size=32, verbose=1)\n",
        "    trained_models.append(model)\n",
        "    group_weights.append(model.get_weights()) # 0 based indexing"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "s4setRtYwzgd",
        "outputId": "b24898c7-00eb-4447-af39-694bd09313b8"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training model for Group 1\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 5s 8ms/step - loss: 7.0067 - accuracy: 0.8271\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 4s 9ms/step - loss: 1.0526 - accuracy: 0.8943\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 4s 8ms/step - loss: 0.5345 - accuracy: 0.9149\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.3782 - accuracy: 0.9303\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2971 - accuracy: 0.9375\n",
            "Training model for Group 2\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 3s 4ms/step - loss: 0.7691 - accuracy: 0.8831\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.3818 - accuracy: 0.9091\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2869 - accuracy: 0.9257\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 3s 6ms/step - loss: 0.2396 - accuracy: 0.9366\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2038 - accuracy: 0.9453\n",
            "Training model for Group 3\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.4633 - accuracy: 0.9131\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2826 - accuracy: 0.9301\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2250 - accuracy: 0.9412\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2022 - accuracy: 0.9459\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 3s 5ms/step - loss: 0.1959 - accuracy: 0.9503\n",
            "Training model for Group 4\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.3312 - accuracy: 0.9307\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2361 - accuracy: 0.9429\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2138 - accuracy: 0.9487\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1747 - accuracy: 0.9555\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1589 - accuracy: 0.9558\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "total_samples = sum(len(split_labels[i]) for i in range(len(split_labels)))"
      ],
      "metadata": {
        "id": "X5-7eoAzn5Mz"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(group_weights[3])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Yqvqdo5n-mV",
        "outputId": "2b49fa01-f715-47ca-b611-1cf706486e18"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "averaged_weights = []\n",
        "for layer_weights in zip(*group_weights):\n",
        "    weighted_sum = np.sum([layer_weights[i] * len(split_labels[i]) for i in range(len(split_labels))], axis=0)\n",
        "    averaged_weight = weighted_sum / total_samples\n",
        "    averaged_weights.append(averaged_weight)"
      ],
      "metadata": {
        "id": "3LjcR1D55xfZ"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.set_weights(averaged_weights)"
      ],
      "metadata": {
        "id": "hbt9GAcx51QM"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_images = tf.convert_to_tensor(test_images.numpy(), dtype=tf.float32)\n",
        "test_labels = tf.convert_to_tensor(test_labels.numpy(), dtype=tf.int32)"
      ],
      "metadata": {
        "id": "kDhVO16L6d-n"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Evaluating the averaged model\")\n",
        "evaluation = model.evaluate(test_images, test_labels, verbose=1)\n",
        "print(f\"Loss: {evaluation[0]}, Accuracy: {evaluation[1]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u-dSPyxN52f1",
        "outputId": "d4a72af8-53d3-4d7c-b5d9-596fba59a429"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Evaluating the averaged model\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2947 - accuracy: 0.9438\n",
            "Loss: 0.29471510648727417, Accuracy: 0.9437999725341797\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Shared Seeding\n"
      ],
      "metadata": {
        "id": "ruv59tWw7nnX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "seed = np.random.randint(0, 1000)\n",
        "tf.random.set_seed(seed)"
      ],
      "metadata": {
        "id": "Bari-kds0RrE"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "\n",
        "trained_models = []\n",
        "group_weights = []  # List to store the weights of each group\n",
        "\n",
        "for i in range(len(split_images)):\n",
        "    print(f\"Training model for Group {i+1}\")\n",
        "    model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "    model.fit(split_images[i], split_labels[i], epochs=5, batch_size=32, verbose=1)\n",
        "    trained_models.append(model)\n",
        "    group_weights.append(model.get_weights()) # 0 based indexing"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3jHZySUa0VRO",
        "outputId": "a7bdb870-8e4c-44b2-ba03-0ced858c2252"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training model for Group 1\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 3s 4ms/step - loss: 0.3608 - accuracy: 0.9284\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2501 - accuracy: 0.9431\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.2256 - accuracy: 0.9441\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 3s 6ms/step - loss: 0.1830 - accuracy: 0.9537\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1569 - accuracy: 0.9589\n",
            "Training model for Group 2\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.3059 - accuracy: 0.9394\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2064 - accuracy: 0.9503\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1908 - accuracy: 0.9535\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.1744 - accuracy: 0.9593\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 3s 6ms/step - loss: 0.1295 - accuracy: 0.9649\n",
            "Training model for Group 3\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 3s 5ms/step - loss: 0.3535 - accuracy: 0.9400\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2124 - accuracy: 0.9505\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1875 - accuracy: 0.9558\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1637 - accuracy: 0.9587\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1625 - accuracy: 0.9609\n",
            "Training model for Group 4\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 4s 7ms/step - loss: 0.3362 - accuracy: 0.9409\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.1971 - accuracy: 0.9525\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1832 - accuracy: 0.9546\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1818 - accuracy: 0.9556\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1443 - accuracy: 0.9607\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "averaged_weights = []\n",
        "for layer_weights in zip(*group_weights):\n",
        "    weighted_sum = np.sum([layer_weights[i] * len(split_labels[i]) for i in range(len(split_labels))], axis=0)\n",
        "    averaged_weight = weighted_sum / total_samples\n",
        "    averaged_weights.append(averaged_weight)"
      ],
      "metadata": {
        "id": "1SaQZl2z5Pl1"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.set_weights(averaged_weights)"
      ],
      "metadata": {
        "id": "9i1pn2h15VqO"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Evaluating the averaged model\")\n",
        "evaluation = model.evaluate(test_images, test_labels, verbose=1)\n",
        "print(f\"Loss: {evaluation[0]}, Accuracy: {evaluation[1]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rHm_D8Qd5Zcq",
        "outputId": "94f940d3-b11e-4814-a77e-98226875dd55"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Evaluating the averaged model\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2556 - accuracy: 0.9535\n",
            "Loss: 0.25559002161026, Accuracy: 0.953499972820282\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Unequal Division of Data"
      ],
      "metadata": {
        "id": "KhbI_ggZ8lz9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "increments = [5000, 10000, 15000, 20000]\n",
        "split_images = []\n",
        "split_labels = []\n",
        "for i, increment in enumerate(increments):\n",
        "    if i == 0:\n",
        "        split_images.append(train_images[:increment])\n",
        "        split_labels.append(train_labels[:increment])\n",
        "    else:\n",
        "        split_images.append(tf.concat([split_images[i-1], train_images[increments[i-1]:increment]], axis=0))\n",
        "        split_labels.append(tf.concat([split_labels[i-1], train_labels[increments[i-1]:increment]], axis=0))"
      ],
      "metadata": {
        "id": "ew_7_skf8oLz"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(len(split_images)):\n",
        "    print(f\"Group {i+1} - Images: {split_images[i].shape}, Labels: {split_labels[i].shape}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NZOHh-2k-O4A",
        "outputId": "a3eef828-f0f7-4ecb-f484-82389718944b"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Group 1 - Images: (5000, 28, 28), Labels: (5000,)\n",
            "Group 2 - Images: (10000, 28, 28), Labels: (10000,)\n",
            "Group 3 - Images: (15000, 28, 28), Labels: (15000,)\n",
            "Group 4 - Images: (20000, 28, 28), Labels: (20000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Random Seeding"
      ],
      "metadata": {
        "id": "gchKOLJR9TBp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "\n",
        "trained_models = []\n",
        "group_weights = []  # List to store the weights of each group\n",
        "\n",
        "for i in range(len(split_images)):\n",
        "    print(f\"Training model for Group {i+1}\")\n",
        "    seed = np.random.randint(0, 1000)\n",
        "    tf.random.set_seed(seed)\n",
        "    model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "    model.fit(split_images[i], split_labels[i], epochs=5, batch_size=32, verbose=1)\n",
        "    trained_models.append(model)\n",
        "    group_weights.append(model.get_weights()) # 0 based indexing"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N-zDjeRG9SJG",
        "outputId": "a1beed70-7922-49a5-fb5c-21e7bde4e9b8"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training model for Group 1\n",
            "Epoch 1/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.2210 - accuracy: 0.9610\n",
            "Epoch 2/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.1243 - accuracy: 0.9706\n",
            "Epoch 3/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.0878 - accuracy: 0.9780\n",
            "Epoch 4/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.0848 - accuracy: 0.9768\n",
            "Epoch 5/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.0813 - accuracy: 0.9796\n",
            "Training model for Group 2\n",
            "Epoch 1/5\n",
            "313/313 [==============================] - 3s 7ms/step - loss: 0.2093 - accuracy: 0.9647\n",
            "Epoch 2/5\n",
            "313/313 [==============================] - 1s 4ms/step - loss: 0.1317 - accuracy: 0.9712\n",
            "Epoch 3/5\n",
            "313/313 [==============================] - 1s 4ms/step - loss: 0.1180 - accuracy: 0.9701\n",
            "Epoch 4/5\n",
            "313/313 [==============================] - 1s 4ms/step - loss: 0.1375 - accuracy: 0.9702\n",
            "Epoch 5/5\n",
            "313/313 [==============================] - 1s 4ms/step - loss: 0.1510 - accuracy: 0.9702\n",
            "Training model for Group 3\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 3s 4ms/step - loss: 0.2436 - accuracy: 0.9661\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1850 - accuracy: 0.9661\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.1417 - accuracy: 0.9692\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1621 - accuracy: 0.9684\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1345 - accuracy: 0.9710\n",
            "Training model for Group 4\n",
            "Epoch 1/5\n",
            "625/625 [==============================] - 3s 4ms/step - loss: 0.2362 - accuracy: 0.9622\n",
            "Epoch 2/5\n",
            "625/625 [==============================] - 2s 4ms/step - loss: 0.1587 - accuracy: 0.9679\n",
            "Epoch 3/5\n",
            "625/625 [==============================] - 3s 5ms/step - loss: 0.1651 - accuracy: 0.9653\n",
            "Epoch 4/5\n",
            "625/625 [==============================] - 3s 4ms/step - loss: 0.1510 - accuracy: 0.9697\n",
            "Epoch 5/5\n",
            "625/625 [==============================] - 2s 3ms/step - loss: 0.1470 - accuracy: 0.9685\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "num_weights_per_group = [len(weights) for weights in group_weights]\n",
        "print(num_weights_per_group)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9mYKn_6trhwq",
        "outputId": "63b3bc77-f53e-4455-be7f-9c78100f6954"
      },
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[4, 4, 4, 4]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "averaged_weights = []\n",
        "for layer_weights in zip(*group_weights):\n",
        "    weighted_sum = np.sum([layer_weights[i] * len(split_labels[i]) for i in range(len(split_labels))], axis=0)\n",
        "    averaged_weight = weighted_sum / total_samples\n",
        "    averaged_weights.append(averaged_weight)"
      ],
      "metadata": {
        "id": "6614sF8J-TCJ"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.set_weights(averaged_weights)"
      ],
      "metadata": {
        "id": "HGfjIL0O_e5z"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Evaluating the averaged model\")\n",
        "evaluation = model.evaluate(test_images, test_labels, verbose=1)\n",
        "print(f\"Loss: {evaluation[0]}, Accuracy: {evaluation[1]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8O-Yow5i_iYk",
        "outputId": "a9570c35-baea-4c49-8cd8-90815b562978"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Evaluating the averaged model\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.3272 - accuracy: 0.9551\n",
            "Loss: 0.32718124985694885, Accuracy: 0.9550999999046326\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Shared Seeding"
      ],
      "metadata": {
        "id": "uYcjZ9sbsuKG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "seed = np.random.randint(0, 1000)\n",
        "tf.random.set_seed(seed)"
      ],
      "metadata": {
        "id": "XhJpSbDC_r6W"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "\n",
        "trained_models = []\n",
        "group_weights = []  # List to store the weights of each group\n",
        "\n",
        "for i in range(len(split_images)):\n",
        "    print(f\"Training model for Group {i+1}\")\n",
        "    model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "    model.fit(split_images[i], split_labels[i], epochs=5, batch_size=32, verbose=1)\n",
        "    trained_models.append(model)\n",
        "    group_weights.append(model.get_weights()) # 0 based indexing"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RoiZOdul_zdk",
        "outputId": "c5ce7929-a0d3-4a85-ebf4-5db9fc657cbd"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training model for Group 1\n",
            "Epoch 1/5\n",
            "157/157 [==============================] - 1s 5ms/step - loss: 0.0787 - accuracy: 0.9826\n",
            "Epoch 2/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.1186 - accuracy: 0.9768\n",
            "Epoch 3/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.0710 - accuracy: 0.9870\n",
            "Epoch 4/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.1032 - accuracy: 0.9802\n",
            "Epoch 5/5\n",
            "157/157 [==============================] - 1s 4ms/step - loss: 0.0858 - accuracy: 0.9820\n",
            "Training model for Group 2\n",
            "Epoch 1/5\n",
            "313/313 [==============================] - 2s 6ms/step - loss: 0.1591 - accuracy: 0.9730\n",
            "Epoch 2/5\n",
            "313/313 [==============================] - 2s 6ms/step - loss: 0.1026 - accuracy: 0.9798\n",
            "Epoch 3/5\n",
            "313/313 [==============================] - 1s 4ms/step - loss: 0.1142 - accuracy: 0.9769\n",
            "Epoch 4/5\n",
            "313/313 [==============================] - 1s 4ms/step - loss: 0.0974 - accuracy: 0.9788\n",
            "Epoch 5/5\n",
            "313/313 [==============================] - 1s 4ms/step - loss: 0.1006 - accuracy: 0.9814\n",
            "Training model for Group 3\n",
            "Epoch 1/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.2001 - accuracy: 0.9716\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.1210 - accuracy: 0.9749\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.1713 - accuracy: 0.9695\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1128 - accuracy: 0.9763\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 2s 4ms/step - loss: 0.1279 - accuracy: 0.9749\n",
            "Training model for Group 4\n",
            "Epoch 1/5\n",
            "625/625 [==============================] - 3s 4ms/step - loss: 0.2225 - accuracy: 0.9661\n",
            "Epoch 2/5\n",
            "625/625 [==============================] - 3s 4ms/step - loss: 0.1908 - accuracy: 0.9678\n",
            "Epoch 3/5\n",
            "625/625 [==============================] - 3s 5ms/step - loss: 0.1635 - accuracy: 0.9672\n",
            "Epoch 4/5\n",
            "625/625 [==============================] - 2s 4ms/step - loss: 0.1351 - accuracy: 0.9719\n",
            "Epoch 5/5\n",
            "625/625 [==============================] - 2s 4ms/step - loss: 0.1563 - accuracy: 0.9686\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "averaged_weights = []\n",
        "for layer_weights in zip(*group_weights):\n",
        "    weighted_sum = np.sum([layer_weights[i] * len(split_labels[i]) for i in range(len(split_labels))], axis=0)\n",
        "    averaged_weight = weighted_sum / total_samples\n",
        "    averaged_weights.append(averaged_weight)"
      ],
      "metadata": {
        "id": "8x6PJswR_6Gz"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.set_weights(averaged_weights)"
      ],
      "metadata": {
        "id": "nFSzf6YS_9XQ"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Evaluating the averaged model\")\n",
        "evaluation = model.evaluate(test_images, test_labels, verbose=1)\n",
        "print(f\"Loss: {evaluation[0]}, Accuracy: {evaluation[1]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jSHzNeN8AAtl",
        "outputId": "e0dc84e8-8fae-4d45-ac5e-02737015bd6a"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Evaluating the averaged model\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.3742 - accuracy: 0.9547\n",
            "Loss: 0.37419867515563965, Accuracy: 0.9546999931335449\n"
          ]
        }
      ]
    }
  ]
}