deploy: 51da293

_modules/neuralprocesses/architectures/agnp.html

@@ -354,13 +354,13 @@ <h1>Source code for neuralprocesses.architectures.agnp</h1><div class="highlight
 <span class="sd">        width (int, optional): Widths of all intermediate MLPs. Defaults to 512.</span>
 <span class="sd">        nonlinearity (Callable or str, optional): Nonlinearity. Can also be specified</span>
 <span class="sd">            as a string: `&quot;ReLU&quot;` or `&quot;LeakyReLU&quot;`. Defaults to ReLUs.</span>
-<span class="sd">        likelihood (str, optional): Likelihood. Must be one of `&quot;het&quot;` or `&quot;lowrank&quot;`.</span>
-<span class="sd">            Defaults to `&quot;lowrank&quot;`.</span>
+<span class="sd">        likelihood (str, optional): Likelihood. Must be one of `&quot;het&quot;`, `&quot;lowrank&quot;`,</span>
+<span class="sd">            `&quot;spikes-beta&quot;`, or `&quot;bernoulli-gamma&quot;`. Defaults to `&quot;lowrank&quot;`.</span>
 <span class="sd">        num_basis_functions (int, optional): Number of basis functions for the</span>
 <span class="sd">            low-rank likelihood. Defaults to 512.</span>
 <span class="sd">        dim_lv (int, optional): Dimensionality of the latent variable. Defaults to 0.</span>
 <span class="sd">        lv_likelihood (str, optional): Likelihood of the latent variable. Must be one of</span>
-<span class="sd">            `&quot;het&quot;`, `&quot;dense&quot;`, or `&quot;spikes-beta&quot;`. Defaults to `&quot;het&quot;`.</span>
+<span class="sd">            `&quot;het&quot;` or `&quot;dense&quot;`. Defaults to `&quot;het&quot;`.</span>
 <span class="sd">        transform (str or tuple[float, float]): Bijection applied to the</span>
 <span class="sd">            output of the model. This can help deal with positive of bounded data.</span>
 <span class="sd">            Must be either `&quot;positive&quot;`, `&quot;exp&quot;`, `&quot;softplus&quot;`, or</span>

_modules/neuralprocesses/architectures/climate.html

@@ -350,8 +350,8 @@ <h1>Source code for neuralprocesses.architectures.climate</h1><div class="highli
 <span class="sd">            to 128.</span>
 <span class="sd">        lr_deg (float, optional): Resolution of the low-resolution grid. Defaults to</span>
 <span class="sd">            0.75.</span>
-<span class="sd">        likelihood (str, optional): Likelihood. Must be one of `&quot;het&quot;` or `&quot;lowrank&quot;.</span>
-<span class="sd">            Defaults to `&quot;lowrank&quot;`.</span>
+<span class="sd">        likelihood (str, optional): Likelihood. Must be one of `&quot;het&quot;`, `&quot;lowrank&quot;`,</span>
+<span class="sd">            `&quot;spikes-beta&quot;`, or `&quot;bernoulli-gamma&quot;`. Defaults to `&quot;lowrank&quot;`.</span>
 <span class="sd">        dtype (dtype, optional): Data type.</span>
 <span class="sd">    &quot;&quot;&quot;</span>
     <span class="n">mlp_width</span> <span class="o">=</span> <span class="mi">128</span>

_modules/neuralprocesses/architectures/convgnp.html

@@ -473,7 +473,7 @@ <h1>Source code for neuralprocesses.architectures.convgnp</h1><div class="highli
 <span class="sd">        margin (float, optional): Margin of the internal discretisation. Defaults to</span>
 <span class="sd">            0.1.</span>
 <span class="sd">        likelihood (str, optional): Likelihood. Must be one of `&quot;het&quot;`, `&quot;lowrank&quot;`,</span>
-<span class="sd">            or `&quot;spikes-beta&quot;`. Defaults to `&quot;lowrank&quot;`.</span>
+<span class="sd">            `&quot;spikes-beta&quot;`, or `&quot;bernoulli-gamma&quot;`. Defaults to `&quot;lowrank&quot;`.</span>
 <span class="sd">        conv_arch (str, optional): Convolutional architecture to use. Must be one of</span>
 <span class="sd">            `&quot;unet[-res][-sep]&quot;` or `&quot;conv[-res][-sep]&quot;`. Defaults to `&quot;unet&quot;`.</span>
 <span class="sd">        unet_channels (tuple[int], optional): Channels of every layer of the UNet.</span>

_modules/neuralprocesses/architectures/gnp.html

@@ -380,13 +380,13 @@ <h1>Source code for neuralprocesses.architectures.gnp</h1><div class="highlight"
 <span class="sd">        width (int, optional): Widths of all intermediate MLPs. Defaults to 512.</span>
 <span class="sd">        nonlinearity (Callable or str, optional): Nonlinearity. Can also be specified</span>
 <span class="sd">            as a string: `&quot;ReLU&quot;` or `&quot;LeakyReLU&quot;`. Defaults to ReLUs.</span>
-<span class="sd">        likelihood (str, optional): Likelihood. Must be one of `&quot;het&quot;` or `&quot;lowrank&quot;`.</span>
-<span class="sd">            Defaults to `&quot;lowrank&quot;`.</span>
+<span class="sd">        likelihood (str, optional): Likelihood. Must be one of `&quot;het&quot;`, `&quot;lowrank&quot;`,</span>
+<span class="sd">            `&quot;spikes-beta&quot;`, or `&quot;bernoulli-gamma&quot;`. Defaults to `&quot;lowrank&quot;`.</span>
 <span class="sd">        num_basis_functions (int, optional): Number of basis functions for the</span>
 <span class="sd">            low-rank likelihood. Defaults to 512.</span>
 <span class="sd">        dim_lv (int, optional): Dimensionality of the latent variable. Defaults to 0.</span>
 <span class="sd">        lv_likelihood (str, optional): Likelihood of the latent variable. Must be one of</span>
-<span class="sd">            `&quot;het&quot;`, `&quot;dense&quot;`, or `&quot;spikes-beta&quot;`. Defaults to `&quot;het&quot;`.</span>
+<span class="sd">            `&quot;het&quot;` or `&quot;dense&quot;`. Defaults to `&quot;het&quot;`.</span>
 <span class="sd">        transform (str or tuple[float, float]): Bijection applied to the</span>
 <span class="sd">            output of the model. This can help deal with positive of bounded data.</span>
 <span class="sd">            Must be either `&quot;positive&quot;`, `&quot;exp&quot;`, `&quot;softplus&quot;`, or</span>

architectures.html

@@ -419,13 +419,13 @@ <h1>List of Predefined Architectures<a class="headerlink" href="#list-of-predefi
 <li><p><strong>width</strong> (<em>int</em><em>, </em><em>optional</em>) – Widths of all intermediate MLPs. Defaults to 512.</p></li>
 <li><p><strong>nonlinearity</strong> (<em>Callable</em><em> or </em><em>str</em><em>, </em><em>optional</em>) – Nonlinearity. Can also be specified
 as a string: <cite>“ReLU”</cite> or <cite>“LeakyReLU”</cite>. Defaults to ReLUs.</p></li>
-<li><p><strong>likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood. Must be one of <cite>“het”</cite> or <cite>“lowrank”</cite>.
-Defaults to <cite>“lowrank”</cite>.</p></li>
+<li><p><strong>likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood. Must be one of <cite>“het”</cite>, <cite>“lowrank”</cite>,
+<cite>“spikes-beta”</cite>, or <cite>“bernoulli-gamma”</cite>. Defaults to <cite>“lowrank”</cite>.</p></li>
 <li><p><strong>num_basis_functions</strong> (<em>int</em><em>, </em><em>optional</em>) – Number of basis functions for the
 low-rank likelihood. Defaults to 512.</p></li>
 <li><p><strong>dim_lv</strong> (<em>int</em><em>, </em><em>optional</em>) – Dimensionality of the latent variable. Defaults to 0.</p></li>
 <li><p><strong>lv_likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood of the latent variable. Must be one of
-<cite>“het”</cite>, <cite>“dense”</cite>, or <cite>“spikes-beta”</cite>. Defaults to <cite>“het”</cite>.</p></li>
+<cite>“het”</cite> or <cite>“dense”</cite>. Defaults to <cite>“het”</cite>.</p></li>
 <li><p><strong>transform</strong> (<em>str</em><em> or </em><em>tuple</em><em>[</em><em>float</em><em>, </em><em>float</em><em>]</em>) – Bijection applied to the
 output of the model. This can help deal with positive of bounded data.
 Must be either <cite>“positive”</cite>, <cite>“exp”</cite>, <cite>“softplus”</cite>, or
@@ -474,13 +474,13 @@ <h1>List of Predefined Architectures<a class="headerlink" href="#list-of-predefi
 <li><p><strong>width</strong> (<em>int</em><em>, </em><em>optional</em>) – Widths of all intermediate MLPs. Defaults to 512.</p></li>
 <li><p><strong>nonlinearity</strong> (<em>Callable</em><em> or </em><em>str</em><em>, </em><em>optional</em>) – Nonlinearity. Can also be specified
 as a string: <cite>“ReLU”</cite> or <cite>“LeakyReLU”</cite>. Defaults to ReLUs.</p></li>
-<li><p><strong>likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood. Must be one of <cite>“het”</cite> or <cite>“lowrank”</cite>.
-Defaults to <cite>“lowrank”</cite>.</p></li>
+<li><p><strong>likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood. Must be one of <cite>“het”</cite>, <cite>“lowrank”</cite>,
+<cite>“spikes-beta”</cite>, or <cite>“bernoulli-gamma”</cite>. Defaults to <cite>“lowrank”</cite>.</p></li>
 <li><p><strong>num_basis_functions</strong> (<em>int</em><em>, </em><em>optional</em>) – Number of basis functions for the
 low-rank likelihood. Defaults to 512.</p></li>
 <li><p><strong>dim_lv</strong> (<em>int</em><em>, </em><em>optional</em>) – Dimensionality of the latent variable. Defaults to 0.</p></li>
 <li><p><strong>lv_likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood of the latent variable. Must be one of
-<cite>“het”</cite>, <cite>“dense”</cite>, or <cite>“spikes-beta”</cite>. Defaults to <cite>“het”</cite>.</p></li>
+<cite>“het”</cite> or <cite>“dense”</cite>. Defaults to <cite>“het”</cite>.</p></li>
 <li><p><strong>transform</strong> (<em>str</em><em> or </em><em>tuple</em><em>[</em><em>float</em><em>, </em><em>float</em><em>]</em>) – Bijection applied to the
 output of the model. This can help deal with positive of bounded data.
 Must be either <cite>“positive”</cite>, <cite>“exp”</cite>, <cite>“softplus”</cite>, or
@@ -527,7 +527,7 @@ <h1>List of Predefined Architectures<a class="headerlink" href="#list-of-predefi
 <li><p><strong>margin</strong> (<em>float</em><em>, </em><em>optional</em>) – Margin of the internal discretisation. Defaults to
 0.1.</p></li>
 <li><p><strong>likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood. Must be one of <cite>“het”</cite>, <cite>“lowrank”</cite>,
-or <cite>“spikes-beta”</cite>. Defaults to <cite>“lowrank”</cite>.</p></li>
+<cite>“spikes-beta”</cite>, or <cite>“bernoulli-gamma”</cite>. Defaults to <cite>“lowrank”</cite>.</p></li>
 <li><p><strong>conv_arch</strong> (<em>str</em><em>, </em><em>optional</em>) – Convolutional architecture to use. Must be one of
 <cite>“unet[-res][-sep]”</cite> or <cite>“conv[-res][-sep]”</cite>. Defaults to <cite>“unet”</cite>.</p></li>
 <li><p><strong>unet_channels</strong> (<em>tuple</em><em>[</em><em>int</em><em>]</em><em>, </em><em>optional</em>) – Channels of every layer of the UNet.
@@ -685,8 +685,8 @@ <h1>List of Predefined Architectures<a class="headerlink" href="#list-of-predefi
 to 128.</p></li>
 <li><p><strong>lr_deg</strong> (<em>float</em><em>, </em><em>optional</em>) – Resolution of the low-resolution grid. Defaults to
 0.75.</p></li>
-<li><p><strong>likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood. Must be one of <cite>“het”</cite> or <cite>“lowrank”.
-Defaults to `”lowrank”</cite>.</p></li>
+<li><p><strong>likelihood</strong> (<em>str</em><em>, </em><em>optional</em>) – Likelihood. Must be one of <cite>“het”</cite>, <cite>“lowrank”</cite>,
+<cite>“spikes-beta”</cite>, or <cite>“bernoulli-gamma”</cite>. Defaults to <cite>“lowrank”</cite>.</p></li>
 <li><p><strong>dtype</strong> (<em>dtype</em><em>, </em><em>optional</em>) – Data type.</p></li>
 </ul>
 </dd>