Tutorial 9: Clarifying qualitative results on train set

jahernando · Oct 17, 2022 · 1f858de · 1f858de
1 parent bae1c8a
commit 1f858de
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 5 deletions.
diff --git a/docs/tutorial_notebooks/JAX/tutorial9/AE_CIFAR10.ipynb b/docs/tutorial_notebooks/JAX/tutorial9/AE_CIFAR10.ipynb
@@ -3326,7 +3326,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "As we initially expected, the reconstruction loss goes down with increasing latent dimensionality. For our model and setup, the two properties seem to be exponentially (or double exponentially) correlated. To understand what these differences in reconstruction error mean, we can visualize example reconstructions of the four models: "
+    "As we initially expected, the reconstruction loss goes down with increasing latent dimensionality. For our model and setup, the two properties seem to be exponentially (or double exponentially) correlated. To understand what these differences in reconstruction error mean, we can visualize example reconstructions of the four models. For simplicity, we visualize four training images of CIFAR10 we have seen already before. For larger models that may overfit, it is recommended to use images from the validation set. "
    ]
   },
   {
@@ -7299,7 +7299,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.5"
+   "version": "3.7.9"
   }
  },
  "nbformat": 4,

diff --git a/docs/tutorial_notebooks/tutorial9/AE_CIFAR10.ipynb b/docs/tutorial_notebooks/tutorial9/AE_CIFAR10.ipynb
@@ -3388,7 +3388,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "As we initially expected, the reconstruction loss goes down with increasing latent dimensionality. For our model and setup, the two properties seem to be exponentially (or double exponentially) correlated. To understand what these differences in reconstruction error mean, we can visualize example reconstructions of the four models: "
+    "As we initially expected, the reconstruction loss goes down with increasing latent dimensionality. For our model and setup, the two properties seem to be exponentially (or double exponentially) correlated. To understand what these differences in reconstruction error mean, we can visualize example reconstructions of the four models. For simplicity, we visualize four training images of CIFAR10 we have seen already before. For larger models that may overfit, it is recommended to use images from the validation set. "
    ]
   },
   {
@@ -7335,7 +7335,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
+   "display_name": "Python 3",
    "language": "python",
    "name": "python3"
   },
@@ -7349,7 +7349,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.2"
+   "version": "3.7.9"
   }
  },
  "nbformat": 4,