sem/CVML-MachineLearning
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

comments

Browse Source
This commit is contained in:
Sem van der Hoeven
2021-06-09 16:20:50 +02:00
parent d07fbe5f2d
commit ac1995feb6
1 changed files with 5 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
exercises/Jupyter notebook CVML.ipynb
View File

@@ -2376,7 +2376,8 @@
"\n", "\n",
"we gebruiken de rmsprop optimizer omdat deze goed te gebruiken is voor kleine\n", "we gebruiken de rmsprop optimizer omdat deze goed te gebruiken is voor kleine\n",
"batch sizes.\n", "batch sizes.\n",
"TODO loss uitleggen\n", "We gebruiken de categorical crossentropy loss function omdat er sprake is van\n",
"meerdere categorieën.\n",
"\"\"\"\n", "\"\"\"\n",
"model.compile(loss='categorical_crossentropy',\n", "model.compile(loss='categorical_crossentropy',\n",
" optimizer='rmsprop',\n", " optimizer='rmsprop',\n",
@@ -2633,6 +2634,8 @@
"compile het model met de binary crossentropy loss function. Dit doen we omdat\n", "compile het model met de binary crossentropy loss function. Dit doen we omdat\n",
"we maar 2 verschillende klassen hebben.\n", "we maar 2 verschillende klassen hebben.\n",
"We gebruiken rmsprop als optimizer omdat we een kleine batch size hebben.\n", "We gebruiken rmsprop als optimizer omdat we een kleine batch size hebben.\n",
"We gebruiken de binary crossentropy loss function omdat we een binaire\n",
"klassificatie hebben.\n",
"Deze optimizer balanceert de step size zodat deze niet te groot of te klein \n", "Deze optimizer balanceert de step size zodat deze niet te groot of te klein \n",
"worden.\n", "worden.\n",
"We gebruiken de accuracy als metric omdat we daarop willen trainen.\n", "We gebruiken de accuracy als metric omdat we daarop willen trainen.\n",
@@ -2767,13 +2770,7 @@
"3/3 [==============================] - 1s 402ms/step - loss: 0.0020 - accuracy: 1.0000 - val_loss: 1.2960 - val_accuracy: 0.8000\n", "3/3 [==============================] - 1s 402ms/step - loss: 0.0020 - accuracy: 1.0000 - val_loss: 1.2960 - val_accuracy: 0.8000\n",
"Epoch 57/100\n", "Epoch 57/100\n",
"3/3 [==============================] - 1s 316ms/step - loss: 1.9244e-05 - accuracy: 1.0000 - val_loss: 1.3116 - val_accuracy: 0.8000\n", "3/3 [==============================] - 1s 316ms/step - loss: 1.9244e-05 - accuracy: 1.0000 - val_loss: 1.3116 - val_accuracy: 0.8000\n",
"Epoch 58/100\n" "Epoch 58/100\n",
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"3/3 [==============================] - 1s 324ms/step - loss: 1.8088e-05 - accuracy: 1.0000 - val_loss: 1.3336 - val_accuracy: 0.8000\n", "3/3 [==============================] - 1s 324ms/step - loss: 1.8088e-05 - accuracy: 1.0000 - val_loss: 1.3336 - val_accuracy: 0.8000\n",
"Epoch 59/100\n", "Epoch 59/100\n",
"3/3 [==============================] - 1s 319ms/step - loss: 3.6851e-05 - accuracy: 1.0000 - val_loss: 1.4207 - val_accuracy: 0.8000\n", "3/3 [==============================] - 1s 319ms/step - loss: 3.6851e-05 - accuracy: 1.0000 - val_loss: 1.4207 - val_accuracy: 0.8000\n",