Fixes parameter counts in tabtransformer.ipynb (#2116)

* Fixes parameter counts in tabtransformer.ipynb

* Fixes parameter counts in tabtransformer.md

* Updates parameter counts in tabtransformer.py

Author: rama100

examples/structured_data/ipynb/tabtransformer.ipynb

@@ -781,7 +781,7 @@
    "source": [
     "The TabTransformer model achieves ~85% validation accuracy.\n",
     "Note that, with the default parameter configurations, both the baseline and the TabTransformer\n",
-    "have similar number of trainable weights: 109,629 and 92,151 respectively, and both use the same training hyperparameters."
+    "have similar number of trainable weights: 109,895 and 87,745 respectively, and both use the same training hyperparameters."
    ]
   },
   {
@@ -831,4 +831,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}

examples/structured_data/md/tabtransformer.md

@@ -690,7 +690,7 @@ Validation accuracy: 84.59%
 </div>
 The TabTransformer model achieves ~85% validation accuracy.
 Note that, with the default parameter configurations, both the baseline and the TabTransformer
-have similar number of trainable weights: 109,629 and 92,151 respectively, and both use the same training hyperparameters.
+have similar number of trainable weights: 109,895 and 87,745 respectively, and both use the same training hyperparameters.
 
 ---
 ## Conclusion

examples/structured_data/tabtransformer.py

@@ -553,7 +553,7 @@ def create_tabtransformer_classifier(
 """
 The TabTransformer model achieves ~85% validation accuracy.
 Note that, with the default parameter configurations, both the baseline and the TabTransformer
-have similar number of trainable weights: 109,629 and 92,151 respectively, and both use the same training hyperparameters.
+have similar number of trainable weights: 109,895 and 87,745 respectively, and both use the same training hyperparameters.
 """
 
 """