Review: key concepts slide move

dlab-berkeley · Apr 22, 2019 · a8ef267 · a8ef267
1 parent 09d77e7
commit a8ef267
Show file tree

Hide file tree

Showing 2 changed files with 76 additions and 76 deletions.
diff --git a/docs/slides.Rmd b/docs/slides.Rmd
@@ -165,44 +165,6 @@ knitr::include_graphics("slide_img/ann.png")
 
 ---
 
-# Key concepts
-.pull-left[
-1) Images are preprocessed 
-- Size
-- Tensor reshape  
-- Image color profile
-
-2) Training and validation sets are split
-- Part 1: MNIST 
-  - Train = 60000, val = 10000
-- Dog-human 
-  - Train = 600, val = 100
-
-3) Define data characteristics
-- Batch size
-- Number of training and validation samples  
-- Epochs
-]
-
-.pull-right[
-4) Define the network
-- Number of units in input, hidden, and output layers. 
-- Activation functions
-- Dropout
-
-5) Compile:  
-- Loss function  
-- Optimizer  
-  - Learning rate
-- Performance metric  
-
-6) Train the model
-
-7) Evaluate
-]
-
----
-
 class: center
 
 # What is an activation function?
@@ -290,6 +252,44 @@ acc = 0.9817
 
 ---
 
+# Review: Key concepts
+.pull-left[
+1) Images are preprocessed 
+- Size
+- Tensor reshape  
+- Image color profile
+
+2) Training and validation sets are split
+- Part 1: MNIST 
+  - Train = 60000, val = 10000
+- Dog-human 
+  - Train = 600, val = 100
+
+3) Define data characteristics
+- Batch size
+- Number of training and validation samples  
+- Epochs
+]
+
+.pull-right[
+4) Define the network
+- Number of units in input, hidden, and output layers. 
+- Activation functions
+- Dropout
+
+5) Compile:  
+- Loss function  
+- Optimizer  
+  - Learning rate
+- Performance metric  
+
+6) Train the model
+
+7) Evaluate
+]
+
+---
+
 # Part 2
 
 Now look at an example that is more realistic in terms of additional preprocessing that you will have to consider if you are using your own images. The MNIST dataset was already curated for us into training and test sets with defined $x$ and $y$ variables. 

diff --git a/docs/slides.html b/docs/slides.html
@@ -129,44 +129,6 @@
 
 ---
 
-# Key concepts
-.pull-left[
-1) Images are preprocessed 
-- Size
-- Tensor reshape  
-- Image color profile
-
-2) Training and validation sets are split
-- Part 1: MNIST 
-  - Train = 60000, val = 10000
-- Dog-human 
-  - Train = 600, val = 100
-
-3) Define data characteristics
-- Batch size
-- Number of training and validation samples  
-- Epochs
-]
-
-.pull-right[
-4) Define the network
-- Number of units in input, hidden, and output layers. 
-- Activation functions
-- Dropout
-
-5) Compile:  
-- Loss function  
-- Optimizer  
-  - Learning rate
-- Performance metric  
-
-6) Train the model
-
-7) Evaluate
-]
-
----
-
 class: center
 
 # What is an activation function?
@@ -248,6 +210,44 @@
 
 ---
 
+# Review: Key concepts
+.pull-left[
+1) Images are preprocessed 
+- Size
+- Tensor reshape  
+- Image color profile
+
+2) Training and validation sets are split
+- Part 1: MNIST 
+  - Train = 60000, val = 10000
+- Dog-human 
+  - Train = 600, val = 100
+
+3) Define data characteristics
+- Batch size
+- Number of training and validation samples  
+- Epochs
+]
+
+.pull-right[
+4) Define the network
+- Number of units in input, hidden, and output layers. 
+- Activation functions
+- Dropout
+
+5) Compile:  
+- Loss function  
+- Optimizer  
+  - Learning rate
+- Performance metric  
+
+6) Train the model
+
+7) Evaluate
+]
+
+---
+
 # Part 2
 
 Now look at an example that is more realistic in terms of additional preprocessing that you will have to consider if you are using your own images. The MNIST dataset was already curated for us into training and test sets with defined `\(x\)` and `\(y\)` variables.