"
]
},
"metadata": {
"needs_background": "light",
"tags": []
},
"output_type": "display_data"
}
],
"source": [
"preds = model.predict(X_test_n[n:n+5])\n",
"f, ax = plt.subplots(1,5)\n",
"f.set_size_inches(80, 40)\n",
"for i,a in enumerate(range(n,n+5)):\n",
" ax[i].imshow(preds[i].reshape(28, 28), cmap='gray')\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {
"colab_type": "text",
"id": "8CIZ2GGXApAL"
},
"source": [
"## Part 2: Variational Autoencoders (VAEs)\n",
"\n",
" Discussion: Why do we care to use VAEs? (i.e., what are the limitations of traditional AEs?)\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"colab_type": "text",
"id": "xcu6XiX9ApAL"
},
"source": [
"### VAE architecture"
]
},
{
"cell_type": "markdown",
"metadata": {
"colab_type": "text",
"id": "n8XI_M3QApAM"
},
"source": [
"\n",
"
![](\"http://www.cvc.uab.es/people/joans/slides_tensorflow/tensorflow_html/vae_files/ae.png\")
\n",
"
![](\"https://raw.githubusercontent.com/Harvard-IACS/2018-CS109A/master/content/styles/iacs.png\")
CS109B Data Science 2: Advanced Topics in Data Science \n",
"## Lab 10: Variational Autoencoders and GANs\n",
"\n",
"**Harvard University**![](\"https://cdn-images-1.medium.com/max/1600/1*ZEvDcg1LP7xvrTSHt0B5-Q@2x.png\")
\n",
"