Deploy to GitHub Pages: 1e92353a

develop/doc/_sources/api/v2/fluid/layers.rst.txt

@@ -305,3 +305,10 @@ lstm_unit
 ---------
 ..  autofunction:: paddle.v2.fluid.layers.lstm_unit
     :noindex:
+
+
+sequence_softmax
+---------
+..  autofunction:: paddle.v2.fluid.layers.sequence_softmax
+    :noindex:
+

develop/doc/api/v2/fluid/layers.html

@@ -1198,6 +1198,47 @@ bias weights will be created and be set to default value.</li>
 </div>
 </dd></dl>
 
+</div>
+<div class="section" id="sequence-softmax">
+<h2>sequence_softmax<a class="headerlink" href="#sequence-softmax" title="Permalink to this headline">¶</a></h2>
+<dl class="function">
+<dt>
+<code class="descclassname">paddle.v2.fluid.layers.</code><code class="descname">sequence_softmax</code><span class="sig-paren">(</span><em>**kwargs</em><span class="sig-paren">)</span></dt>
+<dd><p>Sequence Softmax Operator.</p>
+<p>SequenceSoftmaxOp computes the softmax activation among all time-steps for each
+sequence. The dimension of each time-step should be 1. Thus, the shape of
+input Tensor can be either [N, 1] or [N], where N is the sum of the length
+of all sequences.</p>
+<dl class="docutils">
+<dt>The algorithm works as follows:</dt>
+<dd><dl class="first last docutils">
+<dt>for i-th sequence in a mini-batch:</dt>
+<dd><dl class="first last docutils">
+<dt>$$Out(X[lod[i]:lod[i+1]], :) =</dt>
+<dd>frac{exp(X[lod[i]:lod[i+1], :])}
+{sum(exp(X[lod[i]:lod[i+1], :]))}$$</dd>
+</dl>
+</dd>
+</dl>
+</dd>
+</dl>
+<p>For example, for a mini-batch of 3 sequences with variable-length,
+each containing 2, 3, 2 time-steps, the lod of which is [0, 2, 5, 7],
+then softmax will be computed among X[0:2, :], X[2:5, :], X[5:7, :]
+and N turns out to be 7.</p>
+<table class="docutils field-list" frame="void" rules="none">
+<col class="field-name" />
+<col class="field-body" />
+<tbody valign="top">
+<tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><strong>x</strong> &#8211; (LoDTensor) 1-D or 2-D input LoDTensor with the 2-nd dimension of length 1.
+Duplicable: False  Optional: False</td>
+</tr>
+<tr class="field-even field"><th class="field-name">Returns:</th><td class="field-body">(LoDTensor) 1-D or 2-D output LoDTensor with the 2-nd dimension of length 1.</td>
+</tr>
+</tbody>
+</table>
+</dd></dl>
+
 </div>
 </div>
 

develop/doc_cn/_sources/api/v2/fluid/layers.rst.txt

@@ -305,3 +305,10 @@ lstm_unit
 ---------
 ..  autofunction:: paddle.v2.fluid.layers.lstm_unit
     :noindex:
+
+
+sequence_softmax
+---------
+..  autofunction:: paddle.v2.fluid.layers.sequence_softmax
+    :noindex:
+

develop/doc_cn/api/v2/fluid/layers.html

@@ -1217,6 +1217,47 @@ bias weights will be created and be set to default value.</li>
 </div>
 </dd></dl>
 
+</div>
+<div class="section" id="sequence-softmax">
+<h2>sequence_softmax<a class="headerlink" href="#sequence-softmax" title="永久链接至标题">¶</a></h2>
+<dl class="function">
+<dt>
+<code class="descclassname">paddle.v2.fluid.layers.</code><code class="descname">sequence_softmax</code><span class="sig-paren">(</span><em>**kwargs</em><span class="sig-paren">)</span></dt>
+<dd><p>Sequence Softmax Operator.</p>
+<p>SequenceSoftmaxOp computes the softmax activation among all time-steps for each
+sequence. The dimension of each time-step should be 1. Thus, the shape of
+input Tensor can be either [N, 1] or [N], where N is the sum of the length
+of all sequences.</p>
+<dl class="docutils">
+<dt>The algorithm works as follows:</dt>
+<dd><dl class="first last docutils">
+<dt>for i-th sequence in a mini-batch:</dt>
+<dd><dl class="first last docutils">
+<dt>$$Out(X[lod[i]:lod[i+1]], :) =</dt>
+<dd>frac{exp(X[lod[i]:lod[i+1], :])}
+{sum(exp(X[lod[i]:lod[i+1], :]))}$$</dd>
+</dl>
+</dd>
+</dl>
+</dd>
+</dl>
+<p>For example, for a mini-batch of 3 sequences with variable-length,
+each containing 2, 3, 2 time-steps, the lod of which is [0, 2, 5, 7],
+then softmax will be computed among X[0:2, :], X[2:5, :], X[5:7, :]
+and N turns out to be 7.</p>
+<table class="docutils field-list" frame="void" rules="none">
+<col class="field-name" />
+<col class="field-body" />
+<tbody valign="top">
+<tr class="field-odd field"><th class="field-name">参数:</th><td class="field-body"><strong>x</strong> &#8211; (LoDTensor) 1-D or 2-D input LoDTensor with the 2-nd dimension of length 1.
+Duplicable: False  Optional: False</td>
+</tr>
+<tr class="field-even field"><th class="field-name">返回:</th><td class="field-body">(LoDTensor) 1-D or 2-D output LoDTensor with the 2-nd dimension of length 1.</td>
+</tr>
+</tbody>
+</table>
+</dd></dl>
+
 </div>
 </div>