deploy: 9ff7f01

latest/_sources/autoapi/neural_compressor/adaptor/torch_utils/model_wrapper/index.rst.txt

@@ -33,7 +33,7 @@ Classes
 
 
 
-.. py:class:: TEQLinearFakeQuant(orig_layer, alpha=None, num_bits=4, group_size=-1)
+.. py:class:: TEQLinearFakeQuant(orig_layer, alpha=None, num_bits=4, group_size=-1, scheme='asym')
 
 
 

latest/_sources/autoapi/neural_compressor/adaptor/torch_utils/teq/index.rst.txt

@@ -19,7 +19,7 @@ Classes
 
 
 
-.. py:class:: TEQuantizer(model, weight_config={}, example_inputs=None)
+.. py:class:: TEQuantizer(model, weight_config={}, absorb_to_layer={}, extra_config={}, example_inputs=None)
 
 
    Weight-only quantization, Trainable Equivalent Transformation (TEQ): linear wrapper to apply scale to input

latest/_sources/autoapi/neural_compressor/adaptor/torch_utils/weight_only/index.rst.txt

@@ -210,7 +210,7 @@ Functions
    :rtype: model
 
 
-.. py:function:: teq_quantize(model, weight_config={}, dataloader=None, calib_func=None, example_inputs=None)
+.. py:function:: teq_quantize(model, weight_config={}, absorb_to_layer={}, extra_config={}, dataloader=None, calib_func=None, example_inputs=None)
 
    Run weight-only quantization with
 

latest/autoapi/neural_compressor/adaptor/torch_utils/model_wrapper/index.html

@@ -110,7 +110,7 @@ <h3>Classes<a class="headerlink" href="#classes" title="Permalink to this headin
 
 <dl class="py class">
 <dt class="sig sig-object py" id="neural_compressor.adaptor.torch_utils.model_wrapper.TEQLinearFakeQuant">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">neural_compressor.adaptor.torch_utils.model_wrapper.</span></span><span class="sig-name descname"><span class="pre">TEQLinearFakeQuant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">orig_layer</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_bits</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/intel/neural-compressor/blob/master/neural_compressor/adaptor/torch_utils/model_wrapper.py"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#neural_compressor.adaptor.torch_utils.model_wrapper.TEQLinearFakeQuant" title="Permalink to this definition"></a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">neural_compressor.adaptor.torch_utils.model_wrapper.</span></span><span class="sig-name descname"><span class="pre">TEQLinearFakeQuant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">orig_layer</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_bits</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scheme</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'asym'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/intel/neural-compressor/blob/master/neural_compressor/adaptor/torch_utils/model_wrapper.py"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#neural_compressor.adaptor.torch_utils.model_wrapper.TEQLinearFakeQuant" title="Permalink to this definition"></a></dt>
 <dd><p>wrapper quantization linear</p>
 </dd></dl>
 

latest/autoapi/neural_compressor/adaptor/torch_utils/teq/index.html

@@ -97,7 +97,7 @@ <h3>Classes<a class="headerlink" href="#classes" title="Permalink to this headin
 </table>
 <dl class="py class">
 <dt class="sig sig-object py" id="neural_compressor.adaptor.torch_utils.teq.TEQuantizer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">neural_compressor.adaptor.torch_utils.teq.</span></span><span class="sig-name descname"><span class="pre">TEQuantizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weight_config</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_inputs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/intel/neural-compressor/blob/master/neural_compressor/adaptor/torch_utils/teq.py"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#neural_compressor.adaptor.torch_utils.teq.TEQuantizer" title="Permalink to this definition"></a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">neural_compressor.adaptor.torch_utils.teq.</span></span><span class="sig-name descname"><span class="pre">TEQuantizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weight_config</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">absorb_to_layer</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">extra_config</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_inputs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/intel/neural-compressor/blob/master/neural_compressor/adaptor/torch_utils/teq.py"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#neural_compressor.adaptor.torch_utils.teq.TEQuantizer" title="Permalink to this definition"></a></dt>
 <dd><p>Weight-only quantization, Trainable Equivalent Transformation (TEQ): linear wrapper to apply scale to input</p>
 </dd></dl>
 

latest/autoapi/neural_compressor/adaptor/torch_utils/weight_only/index.html

@@ -114,7 +114,7 @@ <h3>Functions<a class="headerlink" href="#functions" title="Permalink to this he
 <tr class="row-even"><td><p><a class="reference internal" href="#neural_compressor.adaptor.torch_utils.weight_only.awq_quantize" title="neural_compressor.adaptor.torch_utils.weight_only.awq_quantize"><code class="xref py py-obj docutils literal notranslate"><span class="pre">awq_quantize</span></code></a>(model[, weight_config, absorb_dict, ...])</p></td>
 <td><p>Quant the model with Activation-aware Weight quantization(AWQ) method.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#neural_compressor.adaptor.torch_utils.weight_only.teq_quantize" title="neural_compressor.adaptor.torch_utils.weight_only.teq_quantize"><code class="xref py py-obj docutils literal notranslate"><span class="pre">teq_quantize</span></code></a>(model[, weight_config, dataloader, ...])</p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="#neural_compressor.adaptor.torch_utils.weight_only.teq_quantize" title="neural_compressor.adaptor.torch_utils.weight_only.teq_quantize"><code class="xref py py-obj docutils literal notranslate"><span class="pre">teq_quantize</span></code></a>(model[, weight_config, absorb_to_layer, ...])</p></td>
 <td><p>Run weight-only quantization with</p></td>
 </tr>
 </tbody>
@@ -366,7 +366,7 @@ <h3>Functions<a class="headerlink" href="#functions" title="Permalink to this he
 
 <dl class="py function">
 <dt class="sig sig-object py" id="neural_compressor.adaptor.torch_utils.weight_only.teq_quantize">
-<span class="sig-prename descclassname"><span class="pre">neural_compressor.adaptor.torch_utils.weight_only.</span></span><span class="sig-name descname"><span class="pre">teq_quantize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weight_config</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataloader</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">calib_func</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_inputs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/intel/neural-compressor/blob/master/neural_compressor/adaptor/torch_utils/weight_only.py"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#neural_compressor.adaptor.torch_utils.weight_only.teq_quantize" title="Permalink to this definition"></a></dt>
+<span class="sig-prename descclassname"><span class="pre">neural_compressor.adaptor.torch_utils.weight_only.</span></span><span class="sig-name descname"><span class="pre">teq_quantize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weight_config</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">absorb_to_layer</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">extra_config</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataloader</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">calib_func</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_inputs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/intel/neural-compressor/blob/master/neural_compressor/adaptor/torch_utils/weight_only.py"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#neural_compressor.adaptor.torch_utils.weight_only.teq_quantize" title="Permalink to this definition"></a></dt>
 <dd><p>Run weight-only quantization with</p>
 </dd></dl>