Improve instructions

keisen · Jan 31, 2020 · 8e3ff1c · 8e3ff1c
1 parent ba18a48
commit 8e3ff1c
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/examples/visualize_conv_filters.ipynb b/examples/visualize_conv_filters.ipynb
@@ -121,7 +121,7 @@
    "source": [
     "### Visualizing Conv filters\n",
     "\n",
-    "ActivationMaximization will maximize the value that is computed by the loss function that is passed as argument. Here, we try to visualize conv filters. To do so, you need to replace the model output to arbitrary layer's output that contains target conv filters.\n",
+    "ActivationMaximization will maximize the value that is computed by the loss function passed as argument. Here, we try to visualize conv filters. To do so, you need to replace the model output to arbitrary layer's output that contains target conv filters.\n",
     "\n",
     "> [!NOTE]\n",
     "> For visualizing these filters clearly and sharply, we recommend you using `SmoothedLoss`. Most of conv filters will be visualized well."

diff --git a/examples/visualize_dense_layer.ipynb b/examples/visualize_dense_layer.ipynb
@@ -121,7 +121,7 @@
    "source": [
     "### Visualizing a specific output category\n",
     "\n",
-    "ActivationMaximization will maximize the value that is computed by the loss function that is passed as argument. Here, we try to visualize a category as defined No.20 (ouzel) of imagenet.\n",
+    "ActivationMaximization will maximize the value that is computed by the loss function passed as argument. Here, we try to visualize a category as defined No.20 (ouzel) of imagenet.\n",
     "\n",
     "> [!NOTE]\n",
     "> The softmax activation function which is applied to model's last layer may obstruct generating shape images, so that you need to replace the function to a linear function using `model_modifier`."