cnn_2d.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train.min() 56 X_train.max() 220\n",
      "X_test.min() 61 X_test.max() 220\n",
      "y_train.min() 0 y_train.max() 1\n",
      "y_test.min() 0 y_test.max() 1\n",
      "X_train.shape (40, 32, 32)\n",
      "y_train.shape (40, 20)\n",
      "X_test.shape (40, 32, 32)\n",
      "y_test.shape (40, 20)\n",
      "[1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import json\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "from data_utils import flatten_data, normalize_min_max, prepare_data, shuffle_data, normalize_min_max_v2\n",
    "\n",
    "\n",
    "def train_cnn(train_set, test_set):\n",
    "    X_train, y_train = prepare_data(train_set)\n",
    "    X_test, y_test = prepare_data(test_set)\n",
    "\n",
    "    # X_train = normalize_min_max_v2(X_train, -1, 1)\n",
    "    # X_test = normalize_min_max_v2(X_test, -1, 1)\n",
    "\n",
    "    print(\"X_train.min()\", X_train.min(), \"X_train.max()\", X_train.max())\n",
    "    print(\"X_test.min()\", X_test.min(), \"X_test.max()\", X_test.max())\n",
    "    print(\"y_train.min()\", y_train.min(), \"y_train.max()\", y_train.max())\n",
    "    print(\"y_test.min()\", y_test.min(), \"y_test.max()\", y_test.max())\n",
    "\n",
    "    print(\"X_train.shape\", X_train.shape)\n",
    "    print(\"y_train.shape\", y_train.shape)\n",
    "    print(\"X_test.shape\", X_test.shape)\n",
    "    print(\"y_test.shape\", y_test.shape)\n",
    "\n",
    "    # X_train, y_train = shuffle_data(X_train, y_train)\n",
    "\n",
    "    # print(\"X_train.shape\", X_train.shape)\n",
    "    # print(\"y_train.shape\", y_train.shape)\n",
    "    # print(\"X_test.shape\", X_test.shape)\n",
    "    # print(\"y_test.shape\", y_test.shape)\n",
    "\n",
    "    plt.imshow(X_train[0])\n",
    "    print(y_train[0])\n",
    "    plt.show()\n",
    "    plt.imshow(X_test[0])\n",
    "    print(y_test[0])\n",
    "\n",
    "NPY_DATA_DIR = \"npy_datasets\"\n",
    "RESULTS_DIR = \"results\"\n",
    "\n",
    "spectogram_map = {\n",
    "    256: [8, 64, 128, 250],\n",
    "    512: [0, 256, 511],\n",
    "}\n",
    "\n",
    "window_size = 256\n",
    "overlap = 8\n",
    "\n",
    "# Load the dataset\n",
    "test_dataset = np.load(\n",
    "    f\"{NPY_DATA_DIR}/test_{window_size}_{overlap}_png.npy\", allow_pickle=True\n",
    ")\n",
    "retest_dataset = np.load(\n",
    "    f\"{NPY_DATA_DIR}/retest_{window_size}_{overlap}_png.npy\", allow_pickle=True\n",
    ")\n",
    "\n",
    "test_retest_metrics = train_cnn(test_dataset, retest_dataset)\n",
    "# retest_test_metrics = train_cnn(retest_dataset, test_dataset)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "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": "3.7.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}