Tensorflow 2.4.1 has an issue in metrics calculation

tensorflow-2-4-1-has-an-issue-in-metrics-calculation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Tensorflow 2.4.1 has an issue in metrics calculation",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/asterisk37n/9e8c492fc7f01a9f37f9d5fb33ee3406/tensorflow-2-4-1-has-an-issue-in-metrics-calculation.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EZK_OjvBY4Pj",
        "outputId": "2927dfc1-a559-44c2-aa7d-9018cf6e0ae9"
      },
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import random\n",
        "import os\n",
        "\n",
        "print(tf.__version__)\n",
        "\n",
        "tf.random.set_seed(33)\n",
        "os.environ['PYTHONHASHSEED'] = str(33)\n",
        "np.random.seed(33)\n",
        "random.seed(33)\n",
        "\n",
        "model = tf.keras.models.Sequential(\n",
        "    tf.keras.layers.Dense(1, input_shape=(1,))\n",
        ")\n",
        "def my_metric_fn(y_true, y_pred):\n",
        "    squared_difference = tf.square(y_true - y_pred)\n",
        "    loss =  tf.reduce_mean(squared_difference, axis=-1)\n",
        "    tf.print(y_true.shape, y_pred.shape, loss, tf.reduce_mean(squared_difference))\n",
        "    return loss\n",
        "model.compile(optimizer='adam', loss='mean_squared_error', metrics=[my_metric_fn])\n",
        "x = np.random.rand(4,1)\n",
        "y = x ** 2\n",
        "history = model.fit(x=x, y=y, batch_size=2, epochs=1)\n",
        "print(history.history)"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2.4.1\n",
            "TensorShape([2, 1]) TensorShape([2, 1]) [0.216433451 0.167138502] 0.191785976\n",
            "1/2 [==============>...............] - ETA: 0s - loss: 0.1918 - my_metric_fn: 0.1918TensorShape([2, 1]) TensorShape([2, 1]) [0.0477369316 0.0422783121] 0.0450076237\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.1429 - my_metric_fn: 0.1429\n",
            "{'loss': [0.1183968037366867], 'my_metric_fn': [0.1183968037366867]}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ky-fLxMOZbV7",
        "outputId": "facde67f-525d-4138-9204-1b263ae72ac8"
      },
      "source": [
        "(0.191785976 +  0.0450076237)/2 # Where has 0.1429 come from?"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.11839679985"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    }
  ]
}