sd-1.5-square-quantized / INT8 /unet_time_proj.xml
arisha07's picture
Upload 20 files
bf52429
raw
history blame
14.1 kB
<?xml version="1.0" ?>
<net name="torch_jit" version="11">
<layers>
<layer id="64" name="254112541580455" type="Const" version="opset1">
<data offset="943456" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="65" name="254122541671656" type="Const" version="opset1">
<data offset="943460" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="66" name="254132541774728" type="Const" version="opset1">
<data offset="943456" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="67" name="254142541871665" type="Const" version="opset1">
<data offset="943460" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="75" name="cosine_t" type="Parameter" version="opset1">
<data shape="1,160" element_type="f32"/>
<output>
<port id="0" precision="FP32" names="cosine_t">
<dim>1</dim>
<dim>160</dim>
<rt_info/>
</port>
</output>
</layer>
<layer id="81" name="sine_t" type="Parameter" version="opset1">
<data shape="1,160" element_type="f32"/>
<output>
<port id="0" precision="FP32" names="sine_t">
<dim>1</dim>
<dim>160</dim>
<rt_info/>
</port>
</output>
</layer>
<layer id="76" name="/time_proj/Concat/fq_input_1" type="FakeQuantize" version="opset1">
<data levels="256" auto_broadcast="numpy"/>
<input>
<port id="0">
<dim>1</dim>
<dim>160</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1</dim>
</port>
<port id="2">
<dim>1</dim>
<dim>1</dim>
</port>
<port id="3">
<dim>1</dim>
<dim>1</dim>
</port>
<port id="4">
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="5" precision="FP32">
<dim>1</dim>
<dim>160</dim>
</port>
</output>
</layer>
<layer id="77" name="254012540580581" type="Const" version="opset1">
<data offset="943456" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="78" name="254022540678028" type="Const" version="opset1">
<data offset="943460" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="79" name="254032540779540" type="Const" version="opset1">
<data offset="943456" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="80" name="254042540873978" type="Const" version="opset1">
<data offset="943460" size="4" shape="1,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="82" name="/time_proj/Concat/fq_input_0" type="FakeQuantize" version="opset1">
<data levels="256" auto_broadcast="numpy"/>
<input>
<port id="0">
<dim>1</dim>
<dim>160</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1</dim>
</port>
<port id="2">
<dim>1</dim>
<dim>1</dim>
</port>
<port id="3">
<dim>1</dim>
<dim>1</dim>
</port>
<port id="4">
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="5" precision="FP32">
<dim>1</dim>
<dim>160</dim>
</port>
</output>
</layer>
<layer id="83" name="/time_proj/Concat" type="Concat" version="opset1">
<data axis="1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>160</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>160</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="/time_proj/Concat_output_0">
<dim>1</dim>
<dim>320</dim>
</port>
</output>
</layer>
<layer id="84" name="Constant_639583717" type="Const" version="opset1">
<data offset="944112" size="16" shape="2" element_type="i64"/>
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="85" name="Constant_639613718" type="Const" version="opset1">
<data offset="944128" size="16" shape="2" element_type="i64"/>
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="86" name="Constant_639643719" type="Const" version="opset1">
<data offset="944144" size="16" shape="2" element_type="i64"/>
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="87" name="/time_proj/Slice_1" type="StridedSlice" version="opset1">
<data begin_mask="1,0" end_mask="1,0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask=""/>
<input>
<port id="0">
<dim>1</dim>
<dim>320</dim>
</port>
<port id="1">
<dim>2</dim>
</port>
<port id="2">
<dim>2</dim>
</port>
<port id="3">
<dim>2</dim>
</port>
</input>
<output>
<port id="4" precision="FP32" names="/time_proj/Slice_1_output_0">
<dim>1</dim>
<dim>160</dim>
</port>
</output>
</layer>
<layer id="88" name="Constant_639463713" type="Const" version="opset1">
<data offset="944128" size="16" shape="2" element_type="i64"/>
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="89" name="Constant_639493714" type="Const" version="opset1">
<data offset="944160" size="16" shape="2" element_type="i64"/>
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="90" name="Constant_639523715" type="Const" version="opset1">
<data offset="944144" size="16" shape="2" element_type="i64"/>
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="91" name="/time_proj/Slice" type="StridedSlice" version="opset1">
<data begin_mask="1,0" end_mask="1,0" new_axis_mask="" shrink_axis_mask="" ellipsis_mask=""/>
<input>
<port id="0">
<dim>1</dim>
<dim>320</dim>
</port>
<port id="1">
<dim>2</dim>
</port>
<port id="2">
<dim>2</dim>
</port>
<port id="3">
<dim>2</dim>
</port>
</input>
<output>
<port id="4" precision="FP32" names="/time_proj/Slice_output_0">
<dim>1</dim>
<dim>160</dim>
</port>
</output>
</layer>
<layer id="92" name="/time_proj/Concat_1" type="Concat" version="opset1">
<data axis="1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>160</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>160</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="/Cast_output_0,/time_proj/Concat_1_output_0">
<dim>1</dim>
<dim>320</dim>
</port>
</output>
</layer>
<layer id="93" name="time_embedding.linear_1.weight372242271/quantized4854873900" type="Const" version="opset1">
<data offset="944176" size="409600" shape="1280,320" element_type="i8"/>
<output>
<port id="0" precision="I8">
<dim>1280</dim>
<dim>320</dim>
</port>
</output>
</layer>
<layer id="94" name="time_embedding.linear_1.weight372242271/quantized/to_f32" type="Convert" version="opset1">
<data destination_type="f32"/>
<input>
<port id="0">
<dim>1280</dim>
<dim>320</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>1280</dim>
<dim>320</dim>
</port>
</output>
</layer>
<layer id="95" name="/time_embedding/linear_1/Gemm/WithoutBiases/fq_weights_1/zero_point4856772904" type="Const" version="opset1">
<data offset="1353776" size="5120" shape="1280,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1280</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="96" name="/time_embedding/linear_1/Gemm/WithoutBiases/fq_weights_1/minus_zp" type="Subtract" version="opset1">
<data auto_broadcast="numpy"/>
<input>
<port id="0">
<dim>1280</dim>
<dim>320</dim>
</port>
<port id="1">
<dim>1280</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>1280</dim>
<dim>320</dim>
</port>
</output>
</layer>
<layer id="97" name="/time_embedding/linear_1/Gemm/WithoutBiases/fq_weights_1/scale4855681214" type="Const" version="opset1">
<data offset="1358896" size="5120" shape="1280,1" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1280</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="98" name="/time_embedding/linear_1/Gemm/WithoutBiases/fq_weights_1/mulpiply_by_scale" type="Multiply" version="opset1">
<data auto_broadcast="numpy"/>
<input>
<port id="0">
<dim>1280</dim>
<dim>320</dim>
</port>
<port id="1">
<dim>1280</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>1280</dim>
<dim>320</dim>
</port>
</output>
</layer>
<layer id="99" name="/time_embedding/linear_1/Gemm/WithoutBiases" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true"/>
<input>
<port id="0">
<dim>1</dim>
<dim>320</dim>
</port>
<port id="1">
<dim>1280</dim>
<dim>320</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>1</dim>
<dim>1280</dim>
</port>
</output>
</layer>
<layer id="100" name="Constant_74927372472532" type="Const" version="opset1">
<data offset="1364016" size="5120" shape="1,1280" element_type="f32"/>
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1280</dim>
</port>
</output>
</layer>
<layer id="101" name="/time_embedding/linear_1/Gemm" type="Add" version="opset1">
<data auto_broadcast="numpy"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1280</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1280</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="/time_embedding/linear_1/Gemm_output_0">
<dim>1</dim>
<dim>1280</dim>
</port>
</output>
</layer>
<layer id="102" name="_time_embedding_linear_1_Gemm_output0" type="Result" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1280</dim>
</port>
</input>
</layer>
</layers>
<edges>
<edge from-layer="75" from-port="0" to-layer="76" to-port="0"/>
<edge from-layer="64" from-port="0" to-layer="76" to-port="1"/>
<edge from-layer="65" from-port="0" to-layer="76" to-port="2"/>
<edge from-layer="66" from-port="0" to-layer="76" to-port="3"/>
<edge from-layer="67" from-port="0" to-layer="76" to-port="4"/>
<edge from-layer="81" from-port="0" to-layer="82" to-port="0"/>
<edge from-layer="77" from-port="0" to-layer="82" to-port="1"/>
<edge from-layer="78" from-port="0" to-layer="82" to-port="2"/>
<edge from-layer="79" from-port="0" to-layer="82" to-port="3"/>
<edge from-layer="80" from-port="0" to-layer="82" to-port="4"/>
<edge from-layer="76" from-port="5" to-layer="83" to-port="1"/>
<edge from-layer="82" from-port="5" to-layer="83" to-port="0"/>
<edge from-layer="83" from-port="2" to-layer="87" to-port="0"/>
<edge from-layer="83" from-port="2" to-layer="91" to-port="0"/>
<edge from-layer="84" from-port="0" to-layer="87" to-port="1"/>
<edge from-layer="85" from-port="0" to-layer="87" to-port="2"/>
<edge from-layer="86" from-port="0" to-layer="87" to-port="3"/>
<edge from-layer="88" from-port="0" to-layer="91" to-port="1"/>
<edge from-layer="89" from-port="0" to-layer="91" to-port="2"/>
<edge from-layer="90" from-port="0" to-layer="91" to-port="3"/>
<edge from-layer="91" from-port="4" to-layer="92" to-port="0"/>
<edge from-layer="87" from-port="4" to-layer="92" to-port="1"/>
<edge from-layer="92" from-port="2" to-layer="99" to-port="0"/>
<edge from-layer="93" from-port="0" to-layer="94" to-port="0"/>
<edge from-layer="94" from-port="1" to-layer="96" to-port="0"/>
<edge from-layer="95" from-port="0" to-layer="96" to-port="1"/>
<edge from-layer="96" from-port="2" to-layer="98" to-port="0"/>
<edge from-layer="97" from-port="0" to-layer="98" to-port="1"/>
<edge from-layer="98" from-port="2" to-layer="99" to-port="1"/>
<edge from-layer="99" from-port="2" to-layer="101" to-port="0"/>
<edge from-layer="100" from-port="0" to-layer="101" to-port="1"/>
<edge from-layer="101" from-port="2" to-layer="102" to-port="0"/>
</edges>
<rt_info>
<MO_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3"/>
<Runtime_version value="2022.3.0-9052-9752fafe8eb-releases/2022/3"/>
<conversion_parameters>
<framework value="onnx"/>
<input_model value="DIR/unet.onnx"/>
<model_name value="unet"/>
<output_dir value="DIR"/>
</conversion_parameters>
<legacy_frontend value="False"/>
</rt_info>
<quantization_parameters>
<config/>
<version value="{}"/>
<cli_params value="{}"/>
</quantization_parameters>
</net>