New Experimental setup

MelJan · Oct 14, 2018 · 43af13d · 43af13d
1 parent 3eecce2
commit 43af13d
Show file tree

Hide file tree

Showing 5 changed files with 9 additions and 9 deletions.
diff --git a/docs/tutorials/AE_natural_images.rst b/docs/tutorials/AE_natural_images.rst
@@ -38,7 +38,7 @@ The corresponding whitened image patches.
 The learned filters from the whitened natural image patches.
 
 .. figure:: images/SAE_natural_images_filter.png
-   :scale: 10 %
+   :scale: 30 %
    :align: center
    :alt: ICA filter on natural images
 

diff --git a/docs/tutorials/GRBM_natural_images.rst b/docs/tutorials/GRBM_natural_images.rst
@@ -36,7 +36,7 @@ Sampling results for some examples. The first row shows some training data and t
 
 .. figure:: images/GRBM_samples.png
    :alt: GRBM samples
-   :scale: 150 %
+   :scale: 50 %
    :align: center
 
 The log-likelihood and reconstruction error for training and test data

diff --git a/docs/tutorials/ICA_2D_example.rst b/docs/tutorials/ICA_2D_example.rst
@@ -21,7 +21,7 @@ and
 .. raw:: html
 
     <div style="margin-top:10px;">
-      <iframe width="560" height="315" src="http://www.youtube.com/embed/ugiMhRbFnTo" frameborder="0" allowfullscreen></iframe>
+      <iframe width="560" height="315" src="https://www.youtube.com/embed/ugiMhRbFnTo" frameborder="0" allowfullscreen></iframe>
     </div>
 
 Results

diff --git a/docs/tutorials/RBM_MNIST_big.rst b/docs/tutorials/RBM_MNIST_big.rst
@@ -30,7 +30,7 @@ Sampling results for some examples. The first row shows some training data and t
 
 .. figure:: images/BRBM_big_centered_samples.png
    :alt: samples centered
-   :scale: 150 %
+   :scale: 50 %
    :align: center
 
 The log-Likelihood is estimated using annealed importance sampling (optimistic) and reverse annealed importance sampling (pessimistic).
@@ -52,7 +52,7 @@ Sampling results for some examples. The first row shows the training data and th
 
 .. figure:: images/BRBM_big_normal_samples.png
    :alt: samples centered
-   :scale: 150 %
+   :scale: 50 %
    :align: center
 
 .. code-block:: Python

diff --git a/docs/tutorials/RBM_MNIST_small.rst b/docs/tutorials/RBM_MNIST_small.rst
@@ -46,7 +46,7 @@ Sampling results for some examples. The first row shows the training data and th
 
 .. figure:: images/BRBM_small_centered_samples.png
    :alt: samples centered
-   :scale: 150 %
+   :scale: 50 %
    :align: center
 
 The Log-Likelihood is calculated using the exact Partition function, an annealed importance sampling estimation (optimistic) and reverse annealed importance sampling estimation (pessimistic).
@@ -72,7 +72,7 @@ Resulting in the following weights and sampling steps.
 
 .. figure:: images/BRBM_small_normal_samples.png
    :alt: samples normal
-   :scale: 150 %
+   :scale: 50 %
    :align: center
 
 The Log-Likelihood for this model is worse (6.5 nats lower).
@@ -104,7 +104,7 @@ While the centered model has a similar performance on the flipped version,
 
 .. figure:: images/BRBM_small_centered_samples_flipped.png
    :alt: flipped samples centered
-   :scale: 150 %
+   :scale: 50 %
    :align: center
 
 The normal RBM has not.
@@ -122,7 +122,7 @@ The normal RBM has not.
 
 .. figure:: images/BRBM_small_normal_samples_flipped.png
    :alt: flipped samples normal
-   :scale: 150 %
+   :scale: 50 %
    :align: center
 
 For a large number of hidden units see `RBM_MNIST_big <RBM_MNIST_big.html#RBM_MNIST_big>`__.