CircleCI update of dev docs (2963).

Author: Circle CI

master/_downloads/0228827bdcc2a6735b1faf26c864145d/plot_OT_1D_smooth.zip

@@ -139,7 +139,12 @@ <h1>Source code for ot.unbalanced._sinkhorn</h1><div class="highlight"><pre>
 <span class="sd">    - KL is the Kullback-Leibler divergence</span>
 
 <span class="sd">    The algorithm used for solving the problem is the generalized</span>
-<span class="sd">    Sinkhorn-Knopp matrix scaling algorithm as proposed in :ref:`[10, 25] &lt;references-sinkhorn-unbalanced&gt;`</span>
+<span class="sd">    Sinkhorn-Knopp matrix scaling algorithm as proposed in :ref:`[10, 25]</span>
+<span class="sd">    &lt;references-sinkhorn-unbalanced&gt;`</span>
+
+<span class="sd">    .. warning::</span>
+<span class="sd">        Starting from version 0.9.5, the default value has been changed to `reg_type=&#39;kl&#39;` instead of `reg_type=&#39;entropy&#39;`. This makes the function more consistent with the literature</span>
+<span class="sd">        and the other solvers. If you want to use the entropy regularization, please set `reg_type=&#39;entropy&#39;` explicitly.</span>
 
 
 <span class="sd">    Parameters</span>
@@ -175,7 +180,7 @@ <h1>Source code for ot.unbalanced._sinkhorn</h1><div class="highlight"><pre>
 <span class="sd">        + Negative entropy: &#39;entropy&#39;:</span>
 <span class="sd">        :math:`\Omega(\gamma) = \sum_{i,j} \gamma_{i,j} \log(\gamma_{i,j}) - \sum_{i,j} \gamma_{i,j}`.</span>
 <span class="sd">        This is equivalent (up to a constant) to :math:`\Omega(\gamma) = \text{KL}(\gamma, 1_{dim_a} 1_{dim_b}^T)`.</span>
-<span class="sd">        + Kullback-Leibler divergence: &#39;kl&#39;:</span>
+<span class="sd">        + Kullback-Leibler divergence (default): &#39;kl&#39;:</span>
 <span class="sd">        :math:`\Omega(\gamma) = \text{KL}(\gamma, \mathbf{a} \mathbf{b}^T)`.</span>
 <span class="sd">    c : array-like (dim_a, dim_b), optional (default=None)</span>
 <span class="sd">        Reference measure for the regularization.</span>
@@ -368,8 +373,12 @@ <h1>Source code for ot.unbalanced._sinkhorn</h1><div class="highlight"><pre>
 <span class="sd">    - KL is the Kullback-Leibler divergence</span>
 
 <span class="sd">    The algorithm used for solving the problem is the generalized</span>
-<span class="sd">    Sinkhorn-Knopp matrix scaling algorithm as proposed in :ref:`[10, 25] &lt;references-sinkhorn-unbalanced2&gt;`</span>
+<span class="sd">    Sinkhorn-Knopp matrix scaling algorithm as proposed in :ref:`[10, 25]</span>
+<span class="sd">    &lt;references-sinkhorn-unbalanced2&gt;`</span>
 
+<span class="sd">    .. warning::</span>
+<span class="sd">        Starting from version 0.9.5, the default value has been changed to `reg_type=&#39;kl&#39;` instead of `reg_type=&#39;entropy&#39;`. This makes the function more consistent with the literature</span>
+<span class="sd">        and the other solvers. If you want to use the entropy regularization, please set `reg_type=&#39;entropy&#39;` explicitly.</span>
 
 <span class="sd">    Parameters</span>
 <span class="sd">    ----------</span>
@@ -678,6 +687,10 @@ <h1>Source code for ot.unbalanced._sinkhorn</h1><div class="highlight"><pre>
 
 <span class="sd">    The algorithm used for solving the problem is the generalized Sinkhorn-Knopp matrix scaling algorithm as proposed in :ref:`[10, 25] &lt;references-sinkhorn-knopp-unbalanced&gt;`</span>
 
+<span class="sd">    .. warning::</span>
+<span class="sd">        Starting from version 0.9.5, the default value has been changed to `reg_type=&#39;kl&#39;` instead of `reg_type=&#39;entropy&#39;`. This makes the function more consistent with the literature</span>
+<span class="sd">        and the other solvers. If you want to use the entropy regularization, please set `reg_type=&#39;entropy&#39;` explicitly.</span>
+
 
 <span class="sd">    Parameters</span>
 <span class="sd">    ----------</span>
@@ -988,6 +1001,10 @@ <h1>Source code for ot.unbalanced._sinkhorn</h1><div class="highlight"><pre>
 <span class="sd">    log : bool, optional</span>
 <span class="sd">        record `log` if `True`</span>
 
+<span class="sd">    .. warning::</span>
+<span class="sd">        Starting from version 0.9.5, the default value has been changed to `reg_type=&#39;kl&#39;` instead of `reg_type=&#39;entropy&#39;`. This makes the function more consistent with the literature</span>
+<span class="sd">        and the other solvers. If you want to use the entropy regularization, please set `reg_type=&#39;entropy&#39;` explicitly.</span>
+
 
 <span class="sd">    Returns</span>
 <span class="sd">    -------</span>
@@ -1228,7 +1245,6 @@ <h1>Source code for ot.unbalanced._sinkhorn</h1><div class="highlight"><pre>
 
 <span class="sd">    The algorithm used for solving the problem is the translation invariant Sinkhorn algorithm as proposed in :ref:`[73] &lt;references-sinkhorn-unbalanced-translation-invariant&gt;`</span>
 
-
 <span class="sd">    Parameters</span>
 <span class="sd">    ----------</span>
 <span class="sd">    a : array-like (dim_a,)</span>