Exercise MicrosoftDocs#8: 'acc' key should be 'accuracy'

The snippet in exercise MicrosoftDocs#8 is trying to access the key 'acc' that doesn't
exist in keras.models.Sequential().fit(...).history. It should be accessing
'accuracy' instead.

Author: agonzalezro

08. Neural Networks Introduction - Python.ipynb

@@ -82,7 +82,7 @@
         "trusted": true
       },
       "cell_type": "code",
-      "source": "# Let's compile the model\n\n###\n# REPLACE THE <addLoss> WITH 'categorical_crossentropy', <addOptimizer> WITH 'sgd', AND <addMetric> with 'accuracy' (INCLUDING THE QUOTES)\n###\nmodel.compile(loss = <addLoss>, optimizer = <addOptimizer>, metrics = [<addMetric>])\n###\n\n# Time to fit the model\nprint('Starting training')\n\n###\n# REPLACE THE <addTrainX> WITH train_X AND <addTrainY> WITH train_Y\n###\ntraining_stats = model.fit(<addTrainX>, <addTrainY>, batch_size = 1, epochs = 24, verbose = 0)\n###\n\nprint('Training finished')\nprint('Training Evaluation: loss = %0.3f, accuracy = %0.2f%%'\n      %(training_stats.history['loss'][-1], 100 * training_stats.history['acc'][-1]))",
+      "source": "# Let's compile the model\n\n###\n# REPLACE THE <addLoss> WITH 'categorical_crossentropy', <addOptimizer> WITH 'sgd', AND <addMetric> with 'accuracy' (INCLUDING THE QUOTES)\n###\nmodel.compile(loss = <addLoss>, optimizer = <addOptimizer>, metrics = [<addMetric>])\n###\n\n# Time to fit the model\nprint('Starting training')\n\n###\n# REPLACE THE <addTrainX> WITH train_X AND <addTrainY> WITH train_Y\n###\ntraining_stats = model.fit(<addTrainX>, <addTrainY>, batch_size = 1, epochs = 24, verbose = 0)\n###\n\nprint('Training finished')\nprint('Training Evaluation: loss = %0.3f, accuracy = %0.2f%%'\n      %(training_stats.history['loss'][-1], 100 * training_stats.history['accuracy'][-1]))",
       "execution_count": null,
       "outputs": []
     },
@@ -96,7 +96,8 @@
         "trusted": true
       },
       "cell_type": "code",
-      "source": "# Run this!\n\naccuracy, = graph.plot(training_stats.history['acc'],label = 'Accuracy')\ntraining_loss, = graph.plot(training_stats.history['loss'],label = 'Training Loss')\n\n\ngraph.legend(handles = [accuracy,training_loss])\nloss = np.array(training_stats.history['loss'])\nxp = np.linspace(0, loss.shape[0], 10 * loss.shape[0])\ngraph.plot(xp, np.full(xp.shape, 1), c = 'k', linestyle = ':', alpha = 0.5)\ngraph.plot(xp, np.full(xp.shape, 0), c = 'k', linestyle = ':', alpha = 0.5)\ngraph.show()",
+      "source": "# Run this!\n\naccuracy, = graph.plot(training_stats.history['accuracy'],label = '
+      uracy')\ntraining_loss, = graph.plot(training_stats.history['loss'],label = 'Training Loss')\n\n\ngraph.legend(handles = [accuracy,training_loss])\nloss = np.array(training_stats.history['loss'])\nxp = np.linspace(0, loss.shape[0], 10 * loss.shape[0])\ngraph.plot(xp, np.full(xp.shape, 1), c = 'k', linestyle = ':', alpha = 0.5)\ngraph.plot(xp, np.full(xp.shape, 0), c = 'k', linestyle = ':', alpha = 0.5)\ngraph.show()",
       "execution_count": null,
       "outputs": []
     },
@@ -188,4 +189,4 @@
   },
   "nbformat": 4,
   "nbformat_minor": 2
-}
+}