Updating Opacus' website and tutorial to include Ghost Clipping DP-SGD (#667)

Summary:
Pull Request resolved: #667

We update an existing tutorial on fine-tuning a language model, but now also with GC DP-SGD.

Reviewed By: HuanyuZhang

Differential Revision: D61411180

fbshipit-source-id: 515d1a112e5cb9b0ba562128c0a0a7ec7d2a985d

Author: EnayatUllah

tutorials/building_text_classifier.ipynb

@@ -72,6 +72,13 @@ class TutorialHome extends React.Component {
                   Efficient Per-Sample Gradient Computation for More Layers in Opacus
                 </a>
               </li>
+              <li>
+                <a
+                  href="https://pytorch.org/blog/clipping-in-opacus/"
+                  target="_blank">
+                  Enabling Fast Gradient Clipping and Ghost Clipping in Opacus
+                </a>
+              </li>
             </ol>
             <h4>Videos*</h4>
             <p>* Note that Opacus API has changed over time and some of the code samples and demos in the videos may not work. The concepts presented in the videos though are concrete and still valid.</p>

website/pages/tutorials/index.js

@@ -1,12 +1,12 @@
 {
   "Using Opacus": [
     {
-      "id": "building_image_classifier",
-      "title": "Building image classifier with Differential Privacy"
+      "id": "building_text_classifier",
+      "title": "Building text classifier with Fast Gradient Clipping DP-SGD"
     },
     {
-      "id": "building_text_classifier",
-      "title": "Building text classifier with Differential Privacy"
+      "id": "building_image_classifier",
+      "title": "Building image classifier with Differential Privacy"
     },
     {
       "id": "building_lstm_name_classifier",
@@ -27,6 +27,6 @@
     {
       "id": "ddp_tutorial",
       "title": "Training on multiple GPUs with DistributedDataParallel"
-    } 
+    }
   ]
 }