cleaned up some descriptions and added a missing TODO

Author: vyinkabanjo

lab2/solutions/PT_Part2_Debiasing_Solution.ipynb

@@ -1165,7 +1165,7 @@
         "id": "yWCMu12w1BuD"
       },
       "source": [
-        "Now, we will put this decoder together with the standard CNN classifier as our encoder to define the DB-VAE. Note that at this point, there is nothing special about how we put the model together that makes it a \"debiasing\" model -- that will come when we define the training operation. Here, we will define the core VAE architecture by sublassing the `Model` class; defining encoding, reparameterization, and decoding operations; and calling the network end-to-end."
+        "Now, we will put this decoder together with the standard CNN classifier as our encoder to define the DB-VAE. Note that at this point, there is nothing special about how we put the model together that makes it a \"debiasing\" model -- that will come when we define the training operation. Here, we will define the core VAE architecture by sublassing `nn.Module` class; defining encoding, reparameterization, and decoding operations; and calling the network end-to-end."
       ]
     },
     {
@@ -1242,7 +1242,7 @@
         "id": "M-clbYAj2waY"
       },
       "source": [
-        "As stated, the encoder architecture is identical to the CNN from earlier in this lab. Note the outputs of our constructed DB_VAE model in the `call` function: `y_logit, z_mean, z_logsigma, z`. Think carefully about why each of these are outputted and their significance to the problem at hand.\n",
+        "As stated, the encoder architecture is identical to the CNN from earlier in this lab. Note the outputs of our constructed DB_VAE model in the `forward` function: `y_logit, z_mean, z_logsigma, z`. Think carefully about why each of these are outputted and their significance to the problem at hand.\n",
         "\n"
       ]
     },
@@ -1602,10 +1602,13 @@
         "\n",
         "    y_logit, z_mean, z_logsigma, x_recon = dbvae(x)\n",
         "\n",
+        "    '''TODO: call the DB_VAE loss function to compute the loss'''\n",
         "    loss, class_loss = debiasing_loss_function(\n",
         "        x, x_recon, y, y_logit, z_mean, z_logsigma\n",
         "    )\n",
+        "    # loss, class_loss = debiasing_loss_function('''TODO arguments''') # TODO\n",
         "\n",
+        "    '''TODO: backpropagate'''\n",
         "    loss.backward()\n",
         "    optimizer.step()\n",
         "\n",
@@ -1688,8 +1691,6 @@
       ],
       "source": [
         "dbvae.to(device)\n",
-        "\n",
-        "\n",
         "dbvae_logits_list = []\n",
         "for face in test_faces:\n",
         "    face = np.asarray(face, dtype=np.float32)\n",
@@ -1702,7 +1703,6 @@
         "    dbvae_logits_list.append(logit.detach().cpu().numpy())\n",
         "\n",
         "dbvae_logits_array = np.concatenate(dbvae_logits_list, axis=0)\n",
-        "\n",
         "dbvae_logits_tensor = torch.from_numpy(dbvae_logits_array)\n",
         "dbvae_probs_tensor = torch.sigmoid(dbvae_logits_tensor)\n",
         "dbvae_probs_array = dbvae_probs_tensor.squeeze(dim=-1).numpy()\n",
@@ -1712,7 +1712,6 @@
         "std_probs_mean = standard_classifier_probs.mean(axis=1)\n",
         "dbvae_probs_mean = dbvae_probs_array.reshape(len(keys), -1).mean(axis=1)\n",
         "\n",
-        "\n",
         "plt.bar(xx, std_probs_mean, width=0.2, label=\"Standard CNN\")\n",
         "plt.bar(xx + 0.2, dbvae_probs_mean, width=0.2, label=\"DB-VAE\")\n",
         "\n",