CNN Batch Norm & Img Augment: Rem Unused Libs/Funcs

ADolbyB · ADolbyB · commit 18d9a9a8ab16 · 2023-04-14T10:58:14.000-04:00
diff --git a/Lectures/lecture7d.ipynb b/Lectures/lecture7d.ipynb
@@ -17,10 +17,11 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "#### Turn Off Messages:"
+    "#### Turn Off Messages & Set GPU Memory Growth:"
    ]
   },
   {
@@ -31,10 +32,6 @@
    "source": [
     "import os\n",
     "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'\n",
-    "\n",
-    "# Fix: UnknownError: Failed to get convolution algorithm. \n",
-    "# This is probably because cuDNN failed to initialize, \n",
-    "# so try looking to see if a warning log message was printed above\n",
     "os.environ['TF_FORCE_GPU_ALLOW_GROWTH'] = 'true'"
    ]
   },
@@ -100,15 +97,13 @@
     "import numpy as np\n",
     "import pandas as pd\n",
     "from random import randint\n",
-    "from tensorflow.keras import layers, regularizers\n",
     "from keras.models import Sequential, load_model\n",
     "from keras.datasets import cifar10\n",
     "from keras.callbacks import ModelCheckpoint\n",
     "from keras.utils import to_categorical\n",
-    "from keras.applications.vgg16 import preprocess_input\n",
     "from keras.layers import Dense, Conv2D, Flatten, Dropout, Activation, MaxPooling2D, BatchNormalization\n",
     "from keras.preprocessing.image import ImageDataGenerator\n",
-    "from sklearn.metrics import accuracy_score, confusion_matrix, recall_score"
+    "from sklearn.metrics import accuracy_score, confusion_matrix"
    ]
   },
   {
@@ -124,25 +119,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# plot training accuracy and loss\n",
-    "def plotCurve(trainingAccuracy, trainingLoss):\n",
-    "    epochs = np.arange(trainingLoss.shape[0])\n",
-    "    #print(epochs)\n",
-    "    plt.figure(figsize = [12, 6])\n",
-    "    plt.subplot(1, 2, 1)\n",
-    "    plt.plot(epochs, trainingAccuracy)\n",
-    "    #plt.axis([-1, 2, -1, 2])\n",
-    "    plt.xlabel('Epoch')\n",
-    "    plt.ylabel('Accuracy')\n",
-    "    plt.title('Training Accuracy')\n",
-    "\n",
-    "    plt.subplot(1, 2, 2)\n",
-    "    plt.plot(epochs, trainingLoss)\n",
-    "    plt.xlabel('Epoch')\n",
-    "    plt.ylabel('Binary CrossEntropy Loss')\n",
-    "    plt.title('Training Loss')\n",
-    "    plt.show()\n",
-    "\n",
     "# Plot Features of CIFAR-10 data\n",
     "def feat_plot(features, labels, classes, title):\n",
     "    for class_i in classes:\n",
@@ -480,7 +456,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "save_path6 = '../assets/Lecture7/model3_ckpt.h5'\n",
+    "save_path6 = '../assets/Lecture7/model6_ckpt.h5'\n",
     "callbacks_save6 = ModelCheckpoint(save_path6, monitor='val_loss', verbose=0, save_best_only=True, save_freq='epoch')\n",
     "\n",
     "opt6 = tf.keras.optimizers.Adam(learning_rate=0.01)\n",
