v3.ipynb

{
    "cells": [
        {
            "cell_type": "code",
            "execution_count": 1,
            "metadata": {},
            "outputs": [
                {
                    "name": "stderr",
                    "output_type": "stream",
                    "text": [
                        "2024-03-30 09:06:28.323058: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
                        "2024-03-30 09:06:28.348745: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n",
                        "2024-03-30 09:06:28.348767: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n",
                        "2024-03-30 09:06:28.349648: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
                        "2024-03-30 09:06:28.354261: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
                        "To enable the following instructions: AVX2 AVX_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
                        "2024-03-30 09:06:28.787761: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n"
                    ]
                }
            ],
            "source": [
                "import numpy as np\n",
                "import tensorflow as tf\n",
                "from keras.models import Sequential\n",
                "from keras.layers import Dense, Conv2D, MaxPooling2D, Flatten, Dropout, BatchNormalization\n",
                "from keras.losses import categorical_crossentropy\n",
                "from keras.optimizers import Adam\n",
                "import keras\n",
                "\n",
                "from settings import *\n",
                "\n",
                "from utils.general import clean_session, plot_model, split_dataset, scale_dataset, shape_dataset, print_dataset_info\n",
                "from utils.models.ocr import save_model"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 2,
            "metadata": {},
            "outputs": [],
            "source": [
                "import Datasets.OCR.CNN_Letters.loader as CNN_Letters\n",
                "import Datasets.OCR.manual.loader as manual\n",
                "import Datasets.OCR.EnglishFnt.loader as EnglishFnt"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 3,
            "metadata": {},
            "outputs": [],
            "source": [
                "# clean_session()"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 4,
            "metadata": {},
            "outputs": [],
            "source": [
                "def create_model():\n",
                "    model = Sequential()\n",
                "\n",
                "    # CNN\n",
                "    model.add(Conv2D(32, kernel_size=(3, 3), activation='relu',input_shape=INPUT_SHAPE))\n",
                "    model.add(Conv2D(64, kernel_size=(3, 3), activation='relu'))\n",
                "    model.add(Conv2D(128, kernel_size=(3, 3), activation='relu'))\n",
                "\n",
                "    # Pooling\n",
                "    model.add(MaxPooling2D(pool_size=(2, 2), strides=2, padding='same'))\n",
                "\n",
                "    # Drop some data points randomly (mainly for caligraphy reasons)\n",
                "    model.add(Dropout(0.25))\n",
                "\n",
                "    # Flattening before feeding to FCNN\n",
                "    model.add(Flatten())\n",
                "\n",
                "    # FCNN\n",
                "    # use softmax since we are doing classification (1 node for each possible output)\n",
                "    model.add(Dense(128, activation='relu'))\n",
                "    model.add(Dense(64, activation='relu'))\n",
                "    model.add(Dropout(0.3))\n",
                "    model.add(Dense(len(COMBINED_CHARS), activation='softmax'))\n",
                "\n",
                "    model.summary()\n",
                "\n",
                "    return model\n"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 5,
            "metadata": {},
            "outputs": [],
            "source": [
                "def train_model(model, learning_rate, batch_size, epochs, patience):\n",
                "    optimizer = Adam(learning_rate=learning_rate)\n",
                "    model.compile(loss=categorical_crossentropy, optimizer=optimizer, metrics=['accuracy'])\n",
                "\n",
                "    callbacks = [keras.callbacks.EarlyStopping(monitor='val_loss', patience=patience, restore_best_weights=True, verbose=1)]\n",
                "\n",
                "    CX_data, CY_data = CNN_Letters.load_dataset()\n",
                "    MX_data, MY_data = manual.load_dataset()\n",
                "    X_data = np.concatenate((CX_data, MX_data))\n",
                "    Y_data = np.concatenate((CY_data, MY_data))\n",
                "\n",
                "    X_train, X_test, Y_train, Y_test = split_dataset(X_data, Y_data, test_percent=TEST_PERCENT, random_state=RANDOM_STATE, shuffle=True)\n",
                "    X_train, X_test = shape_dataset(X_train, X_test, input_shape=INPUT_SHAPE)\n",
                "    X_train, X_test = scale_dataset(X_train, X_test, batch_size=batch_size)\n",
                "\n",
                "    print_dataset_info(X_data, X_train, Y_train, X_test, Y_test)\n",
                "\n",
                "    history = model.fit(X_train, Y_train, batch_size=batch_size, epochs=epochs, validation_split=VALIDATION_PERCENT, callbacks=callbacks)\n",
                "\n",
                "    return history"
            ]
        },
        {
            "cell_type": "code",
            "execution_count": 6,
            "metadata": {},
            "outputs": [
                {
                    "name": "stdout",
                    "output_type": "stream",
                    "text": [
                        "Model: \"sequential\"\n",
                        "_________________________________________________________________\n",
                        " Layer (type)                Output Shape              Param #   \n",
                        "=================================================================\n",
                        " conv2d (Conv2D)             (None, 98, 48, 32)        320       \n",
                        "                                                                 \n",
                        " conv2d_1 (Conv2D)           (None, 96, 46, 64)        18496     \n",
                        "                                                                 \n",
                        " conv2d_2 (Conv2D)           (None, 94, 44, 128)       73856     \n",
                        "                                                                 \n",
                        " max_pooling2d (MaxPooling2  (None, 47, 22, 128)       0         \n",
                        " D)                                                              \n",
                        "                                                                 \n",
                        " dropout (Dropout)           (None, 47, 22, 128)       0         \n",
                        "                                                                 \n",
                        " flatten (Flatten)           (None, 132352)            0         \n",
                        "                                                                 \n",
                        " dense (Dense)               (None, 128)               16941184  \n",
                        "                                                                 \n",
                        " dense_1 (Dense)             (None, 64)                8256      \n",
                        "                                                                 \n",
                        " dropout_1 (Dropout)         (None, 64)                0         \n",
                        "                                                                 \n",
                        " dense_2 (Dense)             (None, 35)                2275      \n",
                        "                                                                 \n",
                        "=================================================================\n",
                        "Total params: 17044387 (65.02 MB)\n",
                        "Trainable params: 17044387 (65.02 MB)\n",
                        "Non-trainable params: 0 (0.00 Byte)\n",
                        "_________________________________________________________________\n"
                    ]
                },
                {
                    "name": "stderr",
                    "output_type": "stream",
                    "text": [
                        "2024-03-30 09:06:29.908167: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.928139: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.928343: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.929797: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.930001: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.930116: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.974923: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.975062: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.975130: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:901] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
                        "2024-03-30 09:06:29.975187: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1929] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 2170 MB memory:  -> device: 0, name: NVIDIA GeForce RTX 3050 Ti Laptop GPU, pci bus id: 0000:01:00.0, compute capability: 8.6\n"
                    ]
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "bc5bd5944c57424cab04531b17e7da84",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading CNN_Letters:   0%|          | 0/35 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "38e1e700acaa4d1e80b277a8484c869f",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 0:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "0a957afd8ce4444a87edc1fd925da338",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 1:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "e5aef6652ddb4de8a2aa72f869530267",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 2:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "09af5664eb77407c903f5cf30ec6dfa3",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 3:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "6f2cf4ecf7b84bf09047515e599ef6bb",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 4:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "3df2d9db987d4beca3b5d111fdc9f569",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 5:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "0eedc815a6fb48d29ec7bd0be094c6ba",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 6:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "de5286b1b111448d87ee630177371492",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 7:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "b9a04a70746a4578b005aa4ddbe72e50",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 8:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "838ad6b617524825abde5d0a6fbd4c34",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 9:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "72f208ad5cf24686b6716cdfee022c89",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading A:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "b366a10cffa0436784fa58912ef105be",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading B:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "6146b911058d4574a6256ab7517b354f",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading C:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "60ec202e1ef245d48908baf6589ecada",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading D:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "769f2de51b42460291d727a9b0ee0a42",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading E:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "1d99db20be37427cb0c04761cc773234",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading F:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "e1987e1802c448f2a0804c556c1c0f7a",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading G:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "a0c2f6234230460896a09d4d5b51f532",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading H:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "80f3dbf40c854a07b1957e82b40d6d1f",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading I:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "75802ca9e964401aadeca13564d62586",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading J:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "3e5454d7534a4e47bf58a231202666d7",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading K:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "aca82534dbe947fa812820d22f435c72",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading L:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "a350d9792f75463a98e1d30f6acabf7b",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading M:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "9a73cad1ccf5443b975bad3d5f93c4f2",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading N:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "fc5fc2c1820d48eba465c873ec6452ea",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading P:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "37692eb5243a401ab3f7954190be3da4",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading Q:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "1d6677ffd6084a78a5c7929d22b58ba0",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading R:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "0e03edd6a742461387cc6ab7d381bb61",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading S:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "11a769167fee40fcbb99327a770b7d42",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading T:   0%|          | 0/1020 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "0479890632f0489e95f7fca7463d97d2",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading U:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "8c4ccaa0919e48d6b8079ee257970b48",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading V:   0%|          | 0/1030 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "f95dc46735fb420fa9f2c52f4751102c",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading W:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "181ff565ae3a48d28145600a822693b8",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading X:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "814daaa939a4401789c7e117f829b242",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading Y:   0%|          | 0/1010 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "8c5d520decf645b893bd376bcd795ea8",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading Z:   0%|          | 0/810 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "0c391ff5804f4cfd9d4524ebbfdca42a",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading manual:   0%|          | 0/35 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "b88986fd2a844db58412d29199d4a2de",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 0:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "38ae9b70c6e446239094b1c6145d1742",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 1:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "75e33a9c5674458581244fc2fbd0302f",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 2:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "a75b3e9cf5154838b9b1d5819c79ef4e",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 3:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "4a303e82b80d40b7ac11ca1b793257b7",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 4:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "c18ae67c04514b0c8b8f824dd9a09e5a",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 5:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "ebc5c7530774471c86f910bfe6606e56",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 6:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "604df599872348bab608670d3a8fdb44",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 7:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "262663c9a7594d899859c7d1e3974011",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 8:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "7559b09fa1394d09b6fbacc1ec063736",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading 9:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "dacac0a064aa418285d34489db0f4472",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading A:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "cefef8be7b8f495bba8a24e1e7d764ee",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading B:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "f73c9448171c41528054978437cf411c",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading C:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "8d465f37a71445fdbea5006485213a82",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading D:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "426460e4e71246f59875b32896af9704",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading E:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "b42338ef23874b3088a9d0648a0c5523",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading F:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "db9386b8e82b45e5be7cc609fa795111",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading G:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "7540c9f106a54040aeac1eb90b4eec6d",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading H:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "0ccd5a2753a94ec49d4d93c2fb0899e0",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading I:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "db6e440670a640d0ba5cc29ae26ec12c",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading J:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "c53f4f2c645444f682ef1bebaad88e60",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading K:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "01924a11deef48768324fdd85fab054d",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading L:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "5fcb9381e2234e41ad66d6dddeba787c",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading M:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "25f61f7d491d475d94c8081c891b9683",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading N:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "0d0c4d9a43da46d796bf2fba0315dfcc",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading P:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "7cd38483946342d499968780d2b53443",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading Q:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "fca0998bd0224f079e68e52572c38445",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading R:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "6b6047b27cbe4845ae3cf83435f94ac2",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading S:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "2c62a54ab8a44d14bea9579883ec8334",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading T:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "b5a899eda32b4ec0afed3ab545a5f66a",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading U:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "46edaf9228e241dbb2f1202aa1ab688f",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading V:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "d6ed9b8f88fa4e76b3ad5279045473d9",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading W:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "6d4fc29ca905423485fbe3cb98191c25",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading X:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "e32349d2d7544be1a7c759ab77cd5dcf",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading Y:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "data": {
                        "application/vnd.jupyter.widget-view+json": {
                            "model_id": "ea1123ec568442a5a8ca9124852d9f32",
                            "version_major": 2,
                            "version_minor": 0
                        },
                        "text/plain": [
                            "Loading Z:   0%|          | 0/30 [00:00<?, ?it/s]"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "name": "stdout",
                    "output_type": "stream",
                    "text": [
                        "Data samples: 107550\n",
                        "X_train shape: 96795\n",
                        "Y_train shape: 96795\n",
                        "X_test shape: 10755\n",
                        "Y_test shape: 10755\n",
                        "Epoch 1/25\n"
                    ]
                },
                {
                    "name": "stderr",
                    "output_type": "stream",
                    "text": [
                        "2024-03-30 09:06:41.653686: E tensorflow/core/grappler/optimizers/meta_optimizer.cc:961] layout failed: INVALID_ARGUMENT: Size of values 0 does not match size of permutation 4 @ fanin shape insequential/dropout/dropout/SelectV2-2-TransposeNHWCToNCHW-LayoutOptimizer\n",
                        "2024-03-30 09:06:42.080209: I external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:454] Loaded cuDNN version 8906\n",
                        "2024-03-30 09:06:42.136818: I external/local_tsl/tsl/platform/default/subprocess.cc:304] Start cannot spawn child process: No such file or directory\n",
                        "2024-03-30 09:06:42.403671: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 1.03GiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.548014: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 1.03GiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.597492: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 908.12MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.615311: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 412.78MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.615341: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 412.78MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.632020: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 412.78MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.703275: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 412.78MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.747743: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 412.78MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.747766: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 412.78MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:42.771172: W external/local_tsl/tsl/framework/bfc_allocator.cc:296] Allocator (GPU_0_bfc) ran out of memory trying to allocate 412.78MiB with freed_by_count=0. The caller indicates that this is not a failure, but this may mean that there could be performance gains if more memory were available.\n",
                        "2024-03-30 09:06:43.117720: I external/local_tsl/tsl/platform/default/subprocess.cc:304] Start cannot spawn child process: No such file or directory\n",
                        "2024-03-30 09:06:43.399086: I external/local_xla/xla/service/service.cc:168] XLA service 0x7e593cd604b0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n",
                        "2024-03-30 09:06:43.399100: I external/local_xla/xla/service/service.cc:176]   StreamExecutor device (0): NVIDIA GeForce RTX 3050 Ti Laptop GPU, Compute Capability 8.6\n",
                        "2024-03-30 09:06:43.402517: I tensorflow/compiler/mlir/tensorflow/utils/dump_mlir_util.cc:269] disabling MLIR crash reproducer, set env var `MLIR_CRASH_REPRODUCER_DIRECTORY` to enable.\n",
                        "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n",
                        "I0000 00:00:1711782403.451942  197902 device_compiler.h:186] Compiled cluster using XLA!  This line is logged at most once for the lifetime of the process.\n"
                    ]
                },
                {
                    "name": "stdout",
                    "output_type": "stream",
                    "text": [
                        "605/605 [==============================] - 101s 140ms/step - loss: 0.3981 - accuracy: 0.8875 - val_loss: 0.0353 - val_accuracy: 0.9895\n",
                        "Epoch 2/25\n",
                        "605/605 [==============================] - 69s 115ms/step - loss: 0.0518 - accuracy: 0.9844 - val_loss: 0.0168 - val_accuracy: 0.9952\n",
                        "Epoch 3/25\n",
                        "605/605 [==============================] - 69s 115ms/step - loss: 0.0319 - accuracy: 0.9902 - val_loss: 0.0174 - val_accuracy: 0.9949\n",
                        "Epoch 4/25\n",
                        "605/605 [==============================] - 70s 115ms/step - loss: 0.0226 - accuracy: 0.9931 - val_loss: 0.0131 - val_accuracy: 0.9965\n",
                        "Epoch 5/25\n",
                        "605/605 [==============================] - 69s 115ms/step - loss: 0.0177 - accuracy: 0.9945 - val_loss: 0.0094 - val_accuracy: 0.9967\n",
                        "Epoch 6/25\n",
                        "605/605 [==============================] - 68s 113ms/step - loss: 0.0134 - accuracy: 0.9957 - val_loss: 0.0108 - val_accuracy: 0.9978\n",
                        "Epoch 7/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0123 - accuracy: 0.9962 - val_loss: 0.0145 - val_accuracy: 0.9966\n",
                        "Epoch 8/25\n",
                        "605/605 [==============================] - 69s 115ms/step - loss: 0.0118 - accuracy: 0.9964 - val_loss: 0.0110 - val_accuracy: 0.9977\n",
                        "Epoch 9/25\n",
                        "605/605 [==============================] - 69s 115ms/step - loss: 0.0085 - accuracy: 0.9975 - val_loss: 0.0066 - val_accuracy: 0.9985\n",
                        "Epoch 10/25\n",
                        "605/605 [==============================] - 68s 112ms/step - loss: 0.0124 - accuracy: 0.9965 - val_loss: 0.0099 - val_accuracy: 0.9975\n",
                        "Epoch 11/25\n",
                        "605/605 [==============================] - 68s 113ms/step - loss: 0.0065 - accuracy: 0.9978 - val_loss: 0.0075 - val_accuracy: 0.9983\n",
                        "Epoch 12/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0058 - accuracy: 0.9982 - val_loss: 0.0094 - val_accuracy: 0.9980\n",
                        "Epoch 13/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0072 - accuracy: 0.9979 - val_loss: 0.0060 - val_accuracy: 0.9986\n",
                        "Epoch 14/25\n",
                        "605/605 [==============================] - 69s 115ms/step - loss: 0.0119 - accuracy: 0.9968 - val_loss: 0.0099 - val_accuracy: 0.9978\n",
                        "Epoch 15/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0097 - accuracy: 0.9972 - val_loss: 0.0125 - val_accuracy: 0.9971\n",
                        "Epoch 16/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0052 - accuracy: 0.9986 - val_loss: 0.0045 - val_accuracy: 0.9991\n",
                        "Epoch 17/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0048 - accuracy: 0.9985 - val_loss: 0.0053 - val_accuracy: 0.9987\n",
                        "Epoch 18/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0040 - accuracy: 0.9990 - val_loss: 0.0060 - val_accuracy: 0.9987\n",
                        "Epoch 19/25\n",
                        "605/605 [==============================] - 69s 115ms/step - loss: 0.0060 - accuracy: 0.9985 - val_loss: 0.0072 - val_accuracy: 0.9986\n",
                        "Epoch 20/25\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0050 - accuracy: 0.9985 - val_loss: 0.0092 - val_accuracy: 0.9985\n",
                        "Epoch 21/25\n",
                        "605/605 [==============================] - ETA: 0s - loss: 0.0051 - accuracy: 0.9985Restoring model weights from the end of the best epoch: 16.\n",
                        "605/605 [==============================] - 69s 114ms/step - loss: 0.0051 - accuracy: 0.9985 - val_loss: 0.0077 - val_accuracy: 0.9984\n",
                        "Epoch 21: early stopping\n"
                    ]
                },
                {
                    "data": {
                        "image/png": "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",
                        "text/plain": [
                            "<Figure size 1400x500 with 2 Axes>"
                        ]
                    },
                    "metadata": {},
                    "output_type": "display_data"
                },
                {
                    "name": "stdout",
                    "output_type": "stream",
                    "text": [
                        "Saved model v3 to disk\n"
                    ]
                },
                {
                    "ename": "",
                    "evalue": "",
                    "output_type": "error",
                    "traceback": [
                        "\u001b[1;31mThe Kernel crashed while executing code in the current cell or a previous cell. \n",
                        "\u001b[1;31mPlease review the code in the cell(s) to identify a possible cause of the failure. \n",
                        "\u001b[1;31mClick <a href='https://aka.ms/vscodeJupyterKernelCrash'>here</a> for more info. \n",
                        "\u001b[1;31mView Jupyter <a href='command:jupyter.viewOutput'>log</a> for further details."
                    ]
                }
            ],
            "source": [
                "model = create_model()\n",
                "history = train_model(model, learning_rate=0.001, batch_size=128, epochs=25, patience=5)\n",
                "plot_model(history, version=3, save=True)\n",
                "save_model(model, version=3)"
            ]
        }
    ],
    "metadata": {
        "kernelspec": {
            "display_name": "Python 3",
            "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": "3.11.8"
        }
    },
    "nbformat": 4,
    "nbformat_minor": 2
}