ov-tiny-random-gpt2-with-cache / openvino_model.xml
IlyasMoutawwakil's picture
Upload openvino_model.xml with huggingface_hub
d52789a verified
<?xml version="1.0"?>
<net name="Model0" version="11">
<layers>
<layer id="3" name="input_ids" type="Parameter" version="opset1">
<data shape="?,?" element_type="i64" />
<output>
<port id="0" precision="I64" names="input_ids">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="2" name="attention_mask" type="Parameter" version="opset1">
<data shape="?,?" element_type="i64" />
<output>
<port id="0" precision="I64" names="attention_mask">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="1" name="position_ids" type="Parameter" version="opset1">
<data shape="?,?" element_type="i64" />
<output>
<port id="0" precision="I64" names="position_ids">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="0" name="beam_idx" type="Parameter" version="opset1">
<data shape="?" element_type="i32" />
<output>
<port id="0" precision="I32" names="beam_idx">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="4" name="Constant_17159" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="5" name="ShapeOf_17145" type="ShapeOf" version="opset3">
<data output_type="i64" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="6" name="Constant_17147" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="7" name="Constant_17149" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="8" name="Gather_17150" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>2</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="9" name="Constant_17152" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="10" name="Constant_17154" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="11" name="Constant_17156" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="12" name="Concat_17157" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="13" name="Broadcast_17160" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="14" name="ReadValue_16847" type="ReadValue" version="opset6">
<data variable_id="past_key_values.4.valuepresent.4.value" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.4.value">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="15" name="Constant_16531" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="16" name="Gather_16532" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="17" name="self.model.transformer.wte.weight" type="Const" version="opset1">
<data element_type="f32" shape="1000, 32" offset="28" size="128000" />
<output>
<port id="0" precision="FP32" names="self.model.transformer.wte.weight">
<dim>1000</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="18" name="__module.model.transformer/prim::ListConstruct" type="Const" version="opset1">
<data element_type="i64" shape="2" offset="128028" size="16" />
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="19" name="__module.model.transformer/aten::view/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64">
<dim>2</dim>
</port>
</input>
<output>
<port id="2" precision="I64" names="89,input_ids_1">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="20" name="__module.model.transformer.wte/aten::embedding/Convert" type="Convert" version="opset1">
<data destination_type="i32" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="21" name="__module.model.transformer.wte/aten::embedding/Constant" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="22" name="__module.model.transformer.wte/aten::embedding/Gather" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>1000</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="2" precision="I32" />
</input>
<output>
<port id="3" precision="FP32" names="92,inputs_embeds">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="23" name="self.model.transformer.wpe.weight" type="Const" version="opset1">
<data element_type="f32" shape="512, 32" offset="128044" size="65536" />
<output>
<port id="0" precision="FP32" names="self.model.transformer.wpe.weight">
<dim>512</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="24" name="__module.model.transformer.wpe/aten::embedding/Convert" type="Convert" version="opset1">
<data destination_type="i32" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="25" name="__module.model.transformer.wpe/aten::embedding/Constant" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="26" name="__module.model.transformer.wpe/aten::embedding/Gather" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>512</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="2" precision="I32" />
</input>
<output>
<port id="3" precision="FP32" names="101,102,position_embeds,position_embeds.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="27" name="__module.model.transformer/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="103,input.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="28" name="__module.model.transformer.h.0.ln_1/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="29" name="__module.model.transformer.h.0.ln_1/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="156,x.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="30" name="Transpose_13342" type="Const" version="opset1">
<data element_type="f32" shape="96, 32" offset="193584" size="12288" />
<output>
<port id="0" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="31" name="__module.model.transformer.h.0.attn.c_attn/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="166,168,x.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
</output>
</layer>
<layer id="32" name="44" type="Const" version="opset1">
<data element_type="i64" shape="" offset="205872" size="8" />
<output>
<port id="0" precision="I64" names="44" />
</output>
</layer>
<layer id="33" name="__module.model.transformer.h.0.attn/prim::ListUnpack/Concat" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="205880" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="34" name="__module.model.transformer.h.0.attn/prim::ListUnpack/VariadicSplit" type="VariadicSplit" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="3" precision="FP32" names="170,query_states.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="4" precision="FP32" names="171,key_states.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="5" precision="FP32" names="172,value_states.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="35" name="__module.model.transformer.h.0.attn/prim::ListConstruct" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="36" name="__module.model.transformer.h.0.attn/aten::view/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="184">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="37" name="__module.model.transformer.h.0.attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="38" name="__module.model.transformer.h.0.attn/aten::transpose/Transpose" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="185,query.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="39" name="Constant_17171" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="40" name="Constant_17164" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="41" name="Constant_17166" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="42" name="Constant_17168" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="43" name="Concat_17169" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="44" name="Broadcast_17172" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="45" name="ReadValue_16829" type="ReadValue" version="opset6">
<data variable_id="past_key_values.0.keypresent.0.key" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.0.key">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="46" name="Constant_16504" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="47" name="Gather_16505" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="48" name="__module.model.transformer.h.0.attn/prim::ListConstruct_1" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="49" name="__module.model.transformer.h.0.attn/aten::view/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="176">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="50" name="__module.model.transformer.h.0.attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="51" name="__module.model.transformer.h.0.attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="177,key_states.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="52" name="__module.model.transformer.h.0.attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="key.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="53" name="Constant_17183" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="54" name="Constant_17176" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="55" name="Constant_17178" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="56" name="Constant_17180" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="57" name="Concat_17181" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="58" name="Broadcast_17184" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="59" name="ReadValue_16831" type="ReadValue" version="opset6">
<data variable_id="past_key_values.0.valuepresent.0.value" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.0.value">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="60" name="Constant_16507" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="61" name="Gather_16508" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="62" name="Constant_15488" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205952" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="63" name="__module.model.transformer.h.0.attn/aten::view/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="179">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="64" name="__module.model.transformer.h.0.attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="65" name="__module.model.transformer.h.0.attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="180,value_states.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="66" name="__module.model.transformer.h.0.attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="value.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="67" name="__module.model.transformer/aten::new_ones/Broadcast" type="Const" version="opset1">
<data element_type="boolean" shape="" offset="205984" size="1" />
<output>
<port id="0" precision="BOOL" names="132,result.1" />
</output>
</layer>
<layer id="68" name="__module.model.transformer/aten::arange/Constant" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="69" name="ShapeOf_14914" type="ShapeOf" version="opset3">
<data output_type="i64" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="70" name="Constant_14915" type="Const" version="opset1">
<data element_type="i64" shape="" offset="205985" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="71" name="Constant_14916" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="72" name="Gather_14917" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>2</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64" names="95" />
</output>
</layer>
<layer id="73" name="Constant_14918" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="205985" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="74" name="Reshape_14919" type="Reshape" version="opset1">
<data special_zero="false" />
<input>
<port id="0" precision="I64" />
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="75" name="__module.model.transformer/aten::size/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="76" name="__module.model.transformer/aten::size/Gather_3" type="Squeeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I32" />
</input>
<output>
<port id="2" precision="I64" names="107,115" />
</output>
</layer>
<layer id="77" name="ShapeOf_14906" type="ShapeOf" version="opset3">
<data output_type="i64" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="78" name="Constant_14907" type="Const" version="opset1">
<data element_type="i64" shape="" offset="205872" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="79" name="Constant_14908" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="80" name="Gather_14909" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>4</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64" names="109,93" />
</output>
</layer>
<layer id="81" name="__module.model.transformer/aten::add/Add_2" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I64" />
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="111,112,kv_length" />
</output>
</layer>
<layer id="82" name="__module.model.transformer/aten::arange/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="" offset="205993" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="83" name="__module.model.transformer/aten::arange/Range_1" type="Range" version="opset4">
<data output_type="i64" />
<input>
<port id="0" precision="I32" />
<port id="1" precision="I64" />
<port id="2" precision="I32" />
</input>
<output>
<port id="3" precision="I64" names="125">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="84" name="42" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" names="42" />
</output>
</layer>
<layer id="85" name="__module.model.transformer/aten::unsqueeze/Unsqueeze" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="126">
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="86" name="40" type="Const" version="opset1">
<data element_type="i64" shape="" offset="205985" size="8" />
<output>
<port id="0" precision="I64" names="40" />
</output>
</layer>
<layer id="87" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_1" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="127">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="88" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_2" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="128,129,130,kv_idx">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="89" name="__module.model.transformer/aten::le/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="90" name="__module.model.transformer/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I64" />
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="97" />
</output>
</layer>
<layer id="91" name="__module.model.transformer/aten::arange/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="" offset="205993" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="92" name="__module.model.transformer/aten::arange/Range" type="Range" version="opset4">
<data output_type="f32" />
<input>
<port id="0" precision="I64" />
<port id="1" precision="I64" />
<port id="2" precision="I32" />
</input>
<output>
<port id="3" precision="FP32" names="99,cache_position">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="93" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_3" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="116">
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="94" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_4" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="117,118">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="95" name="50" type="Const" version="opset1">
<data element_type="i64" shape="" offset="205997" size="8" />
<output>
<port id="0" precision="I64" names="50" />
</output>
</layer>
<layer id="96" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_5" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="119,q_idx">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="97" name="__module.model.transformer/aten::le/LessEqual" type="LessEqual" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL" names="133,134">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="98" name="__module.model.transformer/aten::__and__/BitwiseAnd" type="BitwiseAnd" version="opset13">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="BOOL" />
<port id="1" precision="BOOL">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL" names="135,result">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="99" name="Constant_15489" type="Const" version="opset1">
<data element_type="i64" shape="2" offset="206005" size="16" />
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="100" name="__module.model.transformer/aten::view/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64">
<dim>2</dim>
</port>
</input>
<output>
<port id="2" precision="I64" names="105,attention_mask.3">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="101" name="__module.model.transformer/aten::to/Convert_2" type="Convert" version="opset1">
<data destination_type="boolean" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="BOOL" names="106,attention_mask.5">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="102" name="__module.model.transformer/aten::index/ShapeOf_1" type="ShapeOf" version="opset3">
<data output_type="i32" />
<input>
<port id="0" precision="BOOL">
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I32">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="103" name="Constant_2928" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="104" name="__module.model.transformer/aten::index/ReduceProd" type="ReduceProd" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="I32">
<dim>2</dim>
</port>
<port id="1" precision="I32" />
</input>
<output>
<port id="2" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="105" name="Constant_2930" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="106" name="__module.model.transformer/aten::index/Concat" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I32">
<dim>1</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="I32">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="107" name="__module.model.transformer/aten::index/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="BOOL">
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>2</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="108" name="__module.model.transformer/aten::index/Convert" type="Convert" version="opset1">
<data destination_type="i32" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="109" name="__module.model.transformer/aten::arange/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="110" name="Constant_14897" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="111" name="Constant_14898" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="112" name="Gather_14899" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>2</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64" names="90">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="113" name="Squeeze_14920" type="Squeeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="1" precision="I64" names="114" />
</output>
</layer>
<layer id="114" name="__module.model.transformer/aten::arange/Constant_4" type="Const" version="opset1">
<data element_type="i32" shape="" offset="205993" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="115" name="__module.model.transformer/aten::arange/Range_2" type="Range" version="opset4">
<data output_type="i64" />
<input>
<port id="0" precision="I32" />
<port id="1" precision="I64" />
<port id="2" precision="I32" />
</input>
<output>
<port id="3" precision="I64" names="120,121">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="116" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_6" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="122">
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="117" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_7" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="123">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="118" name="__module.model.transformer/aten::unsqueeze/Unsqueeze_8" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="124,batch_idx">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="119" name="__module.model.transformer/aten::index/Convert_3" type="Convert" version="opset1">
<data destination_type="i32" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="120" name="__module.model.transformer/aten::index/Constant" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="121" name="__module.model.transformer/aten::index/Split" type="Split" version="opset1">
<data num_splits="2" />
<input>
<port id="0" precision="I32">
<dim>2</dim>
</port>
<port id="1" precision="I32" />
</input>
<output>
<port id="2" precision="I32">
<dim>1</dim>
</port>
<port id="3" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="122" name="__module.model.transformer/aten::index/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I32">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="I32">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="123" name="__module.model.transformer/aten::index/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="I32">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="124" name="__module.model.transformer/aten::index/Gather" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="BOOL">
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="2" precision="I32" />
</input>
<output>
<port id="3" precision="BOOL">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="125" name="Constant_7131" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="206021" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="126" name="__module.model.transformer/aten::index/Reshape_8" type="Reshape" version="opset1">
<data special_zero="false" />
<input>
<port id="0" precision="BOOL">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="127" name="__module.model.transformer/aten::index/ShapeOf_5" type="ShapeOf" version="opset3">
<data output_type="i32" />
<input>
<port id="0" precision="I32">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="128" name="__module.model.transformer/aten::index/Reshape_12" type="Reshape" version="opset1">
<data special_zero="false" />
<input>
<port id="0" precision="BOOL">
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL" names="137,138">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="129" name="__module.model.transformer/aten::__and__/BitwiseAnd_1" type="BitwiseAnd" version="opset13">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="BOOL">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="BOOL">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL" names="139,causal_mask">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="130" name="Constant_11786" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="205985" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="131" name="Constant_216" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="132" name="Unsqueeze_217" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="I64" />
<port id="1" precision="I32" />
</input>
<output>
<port id="2" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="133" name="__module.model.transformer/prim::ListConstruct_2" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="134" name="__module.model.transformer/aten::expand/Broadcast" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="BOOL">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL" names="141,mask">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="135" name="__module.model.transformer/aten::to/Convert_4" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" names="142" />
</output>
</layer>
<layer id="136" name="__module.model.transformer/aten::to/Convert_5" type="Const" version="opset1">
<data element_type="f32" shape="" offset="206029" size="4" />
<output>
<port id="0" precision="FP32" names="144" />
</output>
</layer>
<layer id="137" name="__module.model.transformer/aten::where/Select" type="Select" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="BOOL">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="FP32" />
<port id="2" precision="FP32" />
</input>
<output>
<port id="3" precision="FP32" names="146,191,192,193,304,305,306,417,418,419,530,531,532,643,644,645,attention_mask.7">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="138" name="Constant_489" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="139" name="Constant_14941" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="205985" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="140" name="Reshape_14942" type="Reshape" version="opset1">
<data special_zero="false" />
<input>
<port id="0" precision="I64" />
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="141" name="Add_14945" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="I64" names="190,303,416,529,642">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="142" name="Constant_492" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="205985" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="143" name="Constant_487" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="205997" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="144" name="__module.model.transformer.h.0.attn/aten::slice/Slice_3" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
<port id="4" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="5" precision="FP32" names="194,307,420,533,646,attention_mask.11,attention_mask.13,attention_mask.15,attention_mask.9,attention_mask_1">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="145" name="__module.model.transformer.h.0.attn/aten::scaled_dot_product_attention/ScaledDotProductAttention" type="ScaledDotProductAttention" version="opset13">
<data causal="false" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="4" precision="FP32" names="198,attn_output.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="146" name="__module.model.transformer.h.0.attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="147" name="__module.model.transformer.h.0.attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="199">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="148" name="__module.model.transformer.h.0.attn/prim::ListConstruct_3" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="206033" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="149" name="__module.model.transformer.h.0.attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="204">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="150" name="Transpose_13353" type="Const" version="opset1">
<data element_type="f32" shape="32, 32" offset="206057" size="4096" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="151" name="__module.model.transformer.h.0.attn.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="213,215,218,input.3,x.7">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="152" name="__module.model.transformer.h.0/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="221,input.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="153" name="__module.model.transformer.h.0.ln_2/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="154" name="__module.model.transformer.h.0.ln_2/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="225,x.9">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="155" name="Transpose_13358" type="Const" version="opset1">
<data element_type="f32" shape="128, 32" offset="210153" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="156" name="__module.model.transformer.h.0.mlp.c_fc/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="235,237,input.7,x.11">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="157" name="__module.model.transformer.h.0.mlp.act/aten::mul/Multiply_3" type="Gelu" version="opset7">
<data approximation_mode="TANH" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="245,x.13">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="158" name="Transpose_13372" type="Const" version="opset1">
<data element_type="f32" shape="32, 128" offset="226537" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="159" name="__module.model.transformer.h.0.mlp.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="253,255,input.9,x.15">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="160" name="__module.model.transformer.h.0/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="257,259,input.11">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="161" name="__module.model.transformer.h.1.ln_1/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="162" name="__module.model.transformer.h.1.ln_1/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="269,x.17">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="163" name="Transpose_13377" type="Const" version="opset1">
<data element_type="f32" shape="96, 32" offset="242921" size="12288" />
<output>
<port id="0" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="164" name="__module.model.transformer.h.1.attn.c_attn/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="279,281,x.19">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
</output>
</layer>
<layer id="165" name="__module.model.transformer.h.1.attn/prim::ListUnpack/VariadicSplit" type="VariadicSplit" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="3" precision="FP32" names="283,query_states.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="4" precision="FP32" names="284,key_states.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="5" precision="FP32" names="285,value_states.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="166" name="__module.model.transformer.h.1.attn/prim::ListConstruct" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="167" name="__module.model.transformer.h.1.attn/aten::view/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="297">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="168" name="__module.model.transformer.h.1.attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="169" name="__module.model.transformer.h.1.attn/aten::transpose/Transpose" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="298,query.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="170" name="Constant_17207" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="171" name="Constant_17200" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="172" name="Constant_17202" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="173" name="Constant_17204" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="174" name="Concat_17205" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="175" name="Broadcast_17208" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="176" name="ReadValue_16833" type="ReadValue" version="opset6">
<data variable_id="past_key_values.1.keypresent.1.key" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.1.key">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="177" name="Constant_16510" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="178" name="Gather_16511" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="179" name="__module.model.transformer.h.1.attn/prim::ListConstruct_1" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="180" name="__module.model.transformer.h.1.attn/aten::view/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="289">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="181" name="__module.model.transformer.h.1.attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="182" name="__module.model.transformer.h.1.attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="290,key_states.7">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="183" name="__module.model.transformer.h.1.attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="key.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="184" name="Constant_17195" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="185" name="Constant_17188" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="186" name="Constant_17190" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="187" name="Constant_17192" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="188" name="Concat_17193" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="189" name="Broadcast_17196" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="190" name="ReadValue_16835" type="ReadValue" version="opset6">
<data variable_id="past_key_values.1.valuepresent.1.value" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.1.value">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="191" name="Constant_16513" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="192" name="Gather_16514" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="193" name="Constant_15490" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205952" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="194" name="__module.model.transformer.h.1.attn/aten::view/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="292">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="195" name="__module.model.transformer.h.1.attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="196" name="__module.model.transformer.h.1.attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="293,value_states.7">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="197" name="__module.model.transformer.h.1.attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="value.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="198" name="__module.model.transformer.h.1.attn/aten::scaled_dot_product_attention/ScaledDotProductAttention" type="ScaledDotProductAttention" version="opset13">
<data causal="false" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="4" precision="FP32" names="311,attn_output.7">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="199" name="__module.model.transformer.h.1.attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="200" name="__module.model.transformer.h.1.attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="312">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="201" name="__module.model.transformer.h.1.attn/prim::ListConstruct_3" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="206033" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="202" name="__module.model.transformer.h.1.attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="317">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="203" name="Transpose_13380" type="Const" version="opset1">
<data element_type="f32" shape="32, 32" offset="255209" size="4096" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="204" name="__module.model.transformer.h.1.attn.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="326,328,331,input.13,x.23">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="205" name="__module.model.transformer.h.1/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="334,input.15">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="206" name="__module.model.transformer.h.1.ln_2/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="207" name="__module.model.transformer.h.1.ln_2/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="338,x.25">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="208" name="Transpose_13385" type="Const" version="opset1">
<data element_type="f32" shape="128, 32" offset="259305" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="209" name="__module.model.transformer.h.1.mlp.c_fc/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="348,350,input.17,x.27">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="210" name="__module.model.transformer.h.1.mlp.act/aten::mul/Multiply_3" type="Gelu" version="opset7">
<data approximation_mode="TANH" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="358,x.29">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="211" name="Transpose_13399" type="Const" version="opset1">
<data element_type="f32" shape="32, 128" offset="275689" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="212" name="__module.model.transformer.h.1.mlp.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="366,368,input.19,x.31">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="213" name="__module.model.transformer.h.1/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="370,372,input.21">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="214" name="__module.model.transformer.h.2.ln_1/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="215" name="__module.model.transformer.h.2.ln_1/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="382,x.33">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="216" name="Transpose_13404" type="Const" version="opset1">
<data element_type="f32" shape="96, 32" offset="292073" size="12288" />
<output>
<port id="0" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="217" name="__module.model.transformer.h.2.attn.c_attn/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="392,394,x.35">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
</output>
</layer>
<layer id="218" name="__module.model.transformer.h.2.attn/prim::ListUnpack/VariadicSplit" type="VariadicSplit" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="3" precision="FP32" names="396,query_states.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="4" precision="FP32" names="397,key_states.9">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="5" precision="FP32" names="398,value_states.9">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="219" name="__module.model.transformer.h.2.attn/prim::ListConstruct" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="220" name="__module.model.transformer.h.2.attn/aten::view/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="410">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="221" name="__module.model.transformer.h.2.attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="222" name="__module.model.transformer.h.2.attn/aten::transpose/Transpose" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="411,query.9">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="223" name="Constant_17231" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="224" name="Constant_17224" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="225" name="Constant_17226" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="226" name="Constant_17228" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="227" name="Concat_17229" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="228" name="Broadcast_17232" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="229" name="ReadValue_16837" type="ReadValue" version="opset6">
<data variable_id="past_key_values.2.keypresent.2.key" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.2.key">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="230" name="Constant_16516" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="231" name="Gather_16517" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="232" name="__module.model.transformer.h.2.attn/prim::ListConstruct_1" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="233" name="__module.model.transformer.h.2.attn/aten::view/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="402">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="234" name="__module.model.transformer.h.2.attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="235" name="__module.model.transformer.h.2.attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="403,key_states.11">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="236" name="__module.model.transformer.h.2.attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="key.9">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="237" name="Constant_17219" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="238" name="Constant_17212" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="239" name="Constant_17214" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="240" name="Constant_17216" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="241" name="Concat_17217" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="242" name="Broadcast_17220" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="243" name="ReadValue_16839" type="ReadValue" version="opset6">
<data variable_id="past_key_values.2.valuepresent.2.value" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.2.value">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="244" name="Constant_16519" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="245" name="Gather_16520" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="246" name="Constant_15491" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205952" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="247" name="__module.model.transformer.h.2.attn/aten::view/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="405">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="248" name="__module.model.transformer.h.2.attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="249" name="__module.model.transformer.h.2.attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="406,value_states.11">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="250" name="__module.model.transformer.h.2.attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="value.9">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="251" name="__module.model.transformer.h.2.attn/aten::scaled_dot_product_attention/ScaledDotProductAttention" type="ScaledDotProductAttention" version="opset13">
<data causal="false" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="4" precision="FP32" names="424,attn_output.13">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="252" name="__module.model.transformer.h.2.attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="253" name="__module.model.transformer.h.2.attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="425">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="254" name="__module.model.transformer.h.2.attn/prim::ListConstruct_3" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="206033" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="255" name="__module.model.transformer.h.2.attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="430">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="256" name="Transpose_13407" type="Const" version="opset1">
<data element_type="f32" shape="32, 32" offset="304361" size="4096" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="257" name="__module.model.transformer.h.2.attn.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="439,441,444,input.23,x.39">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="258" name="__module.model.transformer.h.2/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="447,input.25">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="259" name="__module.model.transformer.h.2.ln_2/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="260" name="__module.model.transformer.h.2.ln_2/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="451,x.41">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="261" name="Transpose_13412" type="Const" version="opset1">
<data element_type="f32" shape="128, 32" offset="308457" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="262" name="__module.model.transformer.h.2.mlp.c_fc/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="461,463,input.27,x.43">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="263" name="__module.model.transformer.h.2.mlp.act/aten::mul/Multiply_3" type="Gelu" version="opset7">
<data approximation_mode="TANH" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="471,x.45">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="264" name="Transpose_13426" type="Const" version="opset1">
<data element_type="f32" shape="32, 128" offset="324841" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="265" name="__module.model.transformer.h.2.mlp.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="479,481,input.29,x.47">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="266" name="__module.model.transformer.h.2/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="483,485,input.31">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="267" name="__module.model.transformer.h.3.ln_1/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="268" name="__module.model.transformer.h.3.ln_1/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="495,x.49">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="269" name="Transpose_13431" type="Const" version="opset1">
<data element_type="f32" shape="96, 32" offset="341225" size="12288" />
<output>
<port id="0" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="270" name="__module.model.transformer.h.3.attn.c_attn/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="505,507,x.51">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
</output>
</layer>
<layer id="271" name="__module.model.transformer.h.3.attn/prim::ListUnpack/VariadicSplit" type="VariadicSplit" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="3" precision="FP32" names="509,query_states.7">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="4" precision="FP32" names="510,key_states.13">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="5" precision="FP32" names="511,value_states.13">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="272" name="__module.model.transformer.h.3.attn/prim::ListConstruct" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="273" name="__module.model.transformer.h.3.attn/aten::view/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="523">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="274" name="__module.model.transformer.h.3.attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="275" name="__module.model.transformer.h.3.attn/aten::transpose/Transpose" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="524,query.13">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="276" name="Constant_17255" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="277" name="Constant_17248" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="278" name="Constant_17250" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="279" name="Constant_17252" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="280" name="Concat_17253" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="281" name="Broadcast_17256" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="282" name="ReadValue_16841" type="ReadValue" version="opset6">
<data variable_id="past_key_values.3.keypresent.3.key" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.3.key">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="283" name="Constant_16522" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="284" name="Gather_16523" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="285" name="__module.model.transformer.h.3.attn/prim::ListConstruct_1" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="286" name="__module.model.transformer.h.3.attn/aten::view/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="515">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="287" name="__module.model.transformer.h.3.attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="288" name="__module.model.transformer.h.3.attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="516,key_states.15">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="289" name="__module.model.transformer.h.3.attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="key.13">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="290" name="Constant_17243" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="291" name="Constant_17236" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="292" name="Constant_17238" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="293" name="Constant_17240" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="294" name="Concat_17241" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="295" name="Broadcast_17244" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="296" name="ReadValue_16843" type="ReadValue" version="opset6">
<data variable_id="past_key_values.3.valuepresent.3.value" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.3.value">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="297" name="Constant_16525" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="298" name="Gather_16526" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="299" name="Constant_15492" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205952" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="300" name="__module.model.transformer.h.3.attn/aten::view/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="518">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="301" name="__module.model.transformer.h.3.attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="302" name="__module.model.transformer.h.3.attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="519,value_states.15">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="303" name="__module.model.transformer.h.3.attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="value.13">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="304" name="__module.model.transformer.h.3.attn/aten::scaled_dot_product_attention/ScaledDotProductAttention" type="ScaledDotProductAttention" version="opset13">
<data causal="false" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="4" precision="FP32" names="537,attn_output.19">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="305" name="__module.model.transformer.h.3.attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="306" name="__module.model.transformer.h.3.attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="538">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="307" name="__module.model.transformer.h.3.attn/prim::ListConstruct_3" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="206033" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="308" name="__module.model.transformer.h.3.attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="543">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="309" name="Transpose_13434" type="Const" version="opset1">
<data element_type="f32" shape="32, 32" offset="353513" size="4096" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="310" name="__module.model.transformer.h.3.attn.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="552,554,557,input.33,x.55">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="311" name="__module.model.transformer.h.3/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="560,input.35">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="312" name="__module.model.transformer.h.3.ln_2/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="313" name="__module.model.transformer.h.3.ln_2/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="564,x.57">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="314" name="Transpose_13439" type="Const" version="opset1">
<data element_type="f32" shape="128, 32" offset="357609" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="315" name="__module.model.transformer.h.3.mlp.c_fc/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="574,576,input.37,x.59">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="316" name="__module.model.transformer.h.3.mlp.act/aten::mul/Multiply_3" type="Gelu" version="opset7">
<data approximation_mode="TANH" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="584,x.61">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="317" name="Transpose_13453" type="Const" version="opset1">
<data element_type="f32" shape="32, 128" offset="373993" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="318" name="__module.model.transformer.h.3.mlp.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="592,594,input.39,x.63">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="319" name="__module.model.transformer.h.3/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="596,598,input.41">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="320" name="__module.model.transformer.h.4.ln_1/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="321" name="__module.model.transformer.h.4.ln_1/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="608,x.65">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="322" name="Transpose_13458" type="Const" version="opset1">
<data element_type="f32" shape="96, 32" offset="390377" size="12288" />
<output>
<port id="0" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="323" name="__module.model.transformer.h.4.attn.c_attn/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>96</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="618,620,x.67">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
</output>
</layer>
<layer id="324" name="__module.model.transformer.h.4.attn/prim::ListUnpack/VariadicSplit" type="VariadicSplit" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>96</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="3" precision="FP32" names="622,query_states">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="4" precision="FP32" names="623,key_states.17">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="5" precision="FP32" names="624,value_states.17">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="325" name="Constant_15493" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205952" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="326" name="__module.model.transformer.h.4.attn/aten::view/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="631">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="327" name="__module.model.transformer.h.4.attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="328" name="__module.model.transformer.h.4.attn/aten::transpose/Transpose" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="632,value_states">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="329" name="__module.model.transformer.h.4.attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="present.4.value,value.17">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="331" name="Constant_17267" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="332" name="Constant_17260" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="12" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="333" name="Constant_17262" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="4" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="334" name="Constant_17264" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="20" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="335" name="Concat_17265" type="Concat" version="opset1">
<data axis="0" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
<port id="2" precision="I64">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="336" name="Broadcast_17268" type="Broadcast" version="opset3">
<data mode="numpy" />
<input>
<port id="0" precision="FP32" />
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="337" name="ReadValue_16845" type="ReadValue" version="opset6">
<data variable_id="past_key_values.4.keypresent.4.key" variable_type="f32" variable_shape="?,4,?,8" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>0</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.4.key">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="338" name="Constant_16528" type="Const" version="opset1">
<data element_type="i64" shape="" offset="4" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="339" name="Gather_16529" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="340" name="__module.model.transformer.h.4.attn/prim::ListConstruct" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="341" name="__module.model.transformer.h.4.attn/aten::view/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="628">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="342" name="__module.model.transformer.h.4.attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="343" name="__module.model.transformer.h.4.attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="629,key_states">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="344" name="__module.model.transformer.h.4.attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="key.17">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="354" name="__module.model.transformer.h.4.attn/prim::ListConstruct_2" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="205904" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="355" name="__module.model.transformer.h.4.attn/aten::view/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="636">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="356" name="__module.model.transformer.h.4.attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="357" name="__module.model.transformer.h.4.attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="637,query.17">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="358" name="__module.model.transformer.h.4.attn/aten::scaled_dot_product_attention/ScaledDotProductAttention" type="ScaledDotProductAttention" version="opset13">
<data causal="false" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="4" precision="FP32" names="650,attn_output.25">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="359" name="__module.model.transformer.h.4.attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="205936" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="360" name="__module.model.transformer.h.4.attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="651">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="361" name="__module.model.transformer.h.4.attn/prim::ListConstruct_3" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="206033" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="362" name="__module.model.transformer.h.4.attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>8</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="656">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="363" name="Transpose_13461" type="Const" version="opset1">
<data element_type="f32" shape="32, 32" offset="402665" size="4096" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="364" name="__module.model.transformer.h.4.attn.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="665,667,670,input.43,x.71">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="365" name="__module.model.transformer.h.4/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="673,input.45">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="366" name="__module.model.transformer.h.4.ln_2/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="367" name="__module.model.transformer.h.4.ln_2/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="677,x.73">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="368" name="Transpose_13466" type="Const" version="opset1">
<data element_type="f32" shape="128, 32" offset="406761" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="369" name="__module.model.transformer.h.4.mlp.c_fc/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>128</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="687,689,input.47,x.75">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="370" name="__module.model.transformer.h.4.mlp.act/aten::mul/Multiply_3" type="Gelu" version="opset7">
<data approximation_mode="TANH" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="697,x.77">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="371" name="Transpose_13480" type="Const" version="opset1">
<data element_type="f32" shape="32, 128" offset="423145" size="16384" />
<output>
<port id="0" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="372" name="__module.model.transformer.h.4.mlp.c_proj/aten::addmm/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>128</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="705,707,input.49,x">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="373" name="__module.model.transformer.h.4/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="709,711,input.51">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="374" name="__module.model.transformer.ln_f/aten::layer_norm/Multiply" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="193580" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="375" name="__module.model.transformer.ln_f/aten::layer_norm/MVN" type="MVN" version="opset6">
<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I32">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="717,hidden_states.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="376" name="Constant_15494" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="439529" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="377" name="__module.model.transformer/aten::view/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="719,721,732,733,734,hidden_states,input">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="378" name="__module.model.lm_head/aten::linear/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
<port id="1" precision="FP32">
<dim>1000</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="logits">
<dim>-1</dim>
<dim>-1</dim>
<dim>1000</dim>
</port>
</output>
</layer>
<layer id="353" name="Assign_16830" type="Assign" version="opset6">
<data variable_id="past_key_values.0.keypresent.0.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="352" name="Assign_16832" type="Assign" version="opset6">
<data variable_id="past_key_values.0.valuepresent.0.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="351" name="Assign_16834" type="Assign" version="opset6">
<data variable_id="past_key_values.1.keypresent.1.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="350" name="Assign_16836" type="Assign" version="opset6">
<data variable_id="past_key_values.1.valuepresent.1.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="349" name="Assign_16838" type="Assign" version="opset6">
<data variable_id="past_key_values.2.keypresent.2.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="348" name="Assign_16840" type="Assign" version="opset6">
<data variable_id="past_key_values.2.valuepresent.2.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="347" name="Assign_16842" type="Assign" version="opset6">
<data variable_id="past_key_values.3.keypresent.3.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="346" name="Assign_16844" type="Assign" version="opset6">
<data variable_id="past_key_values.3.valuepresent.3.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="345" name="Assign_16846" type="Assign" version="opset6">
<data variable_id="past_key_values.4.keypresent.4.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="330" name="Assign_16848" type="Assign" version="opset6">
<data variable_id="past_key_values.4.valuepresent.4.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="379" name="Result_2978" type="Result" version="opset1" output_names="logits">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1000</dim>
</port>
</input>
</layer>
</layers>
<edges>
<edge from-layer="0" from-port="0" to-layer="61" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="178" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="192" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="47" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="339" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="298" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="284" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="245" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="16" to-port="1" />
<edge from-layer="0" from-port="0" to-layer="231" to-port="1" />
<edge from-layer="1" from-port="0" to-layer="24" to-port="0" />
<edge from-layer="2" from-port="0" to-layer="100" to-port="0" />
<edge from-layer="3" from-port="0" to-layer="5" to-port="0" />
<edge from-layer="3" from-port="0" to-layer="19" to-port="0" />
<edge from-layer="4" from-port="0" to-layer="13" to-port="0" />
<edge from-layer="5" from-port="1" to-layer="8" to-port="0" />
<edge from-layer="6" from-port="0" to-layer="8" to-port="1" />
<edge from-layer="7" from-port="0" to-layer="8" to-port="2" />
<edge from-layer="8" from-port="3" to-layer="43" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="280" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="241" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="227" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="294" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="335" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="12" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="188" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="174" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="57" to-port="0" />
<edge from-layer="9" from-port="0" to-layer="12" to-port="1" />
<edge from-layer="10" from-port="0" to-layer="12" to-port="2" />
<edge from-layer="11" from-port="0" to-layer="12" to-port="3" />
<edge from-layer="12" from-port="4" to-layer="13" to-port="1" />
<edge from-layer="13" from-port="2" to-layer="14" to-port="0" />
<edge from-layer="14" from-port="1" to-layer="16" to-port="0" />
<edge from-layer="15" from-port="0" to-layer="16" to-port="2" />
<edge from-layer="16" from-port="3" to-layer="329" to-port="0" />
<edge from-layer="17" from-port="0" to-layer="378" to-port="1" />
<edge from-layer="17" from-port="0" to-layer="22" to-port="0" />
<edge from-layer="18" from-port="0" to-layer="19" to-port="1" />
<edge from-layer="19" from-port="2" to-layer="69" to-port="0" />
<edge from-layer="19" from-port="2" to-layer="20" to-port="0" />
<edge from-layer="20" from-port="1" to-layer="22" to-port="1" />
<edge from-layer="21" from-port="0" to-layer="22" to-port="2" />
<edge from-layer="22" from-port="3" to-layer="27" to-port="0" />
<edge from-layer="23" from-port="0" to-layer="26" to-port="0" />
<edge from-layer="24" from-port="1" to-layer="26" to-port="1" />
<edge from-layer="25" from-port="0" to-layer="26" to-port="2" />
<edge from-layer="26" from-port="3" to-layer="27" to-port="1" />
<edge from-layer="27" from-port="2" to-layer="29" to-port="0" />
<edge from-layer="27" from-port="2" to-layer="152" to-port="1" />
<edge from-layer="28" from-port="0" to-layer="29" to-port="1" />
<edge from-layer="29" from-port="2" to-layer="31" to-port="0" />
<edge from-layer="30" from-port="0" to-layer="31" to-port="1" />
<edge from-layer="31" from-port="2" to-layer="34" to-port="0" />
<edge from-layer="32" from-port="0" to-layer="165" to-port="1" />
<edge from-layer="32" from-port="0" to-layer="117" to-port="1" />
<edge from-layer="32" from-port="0" to-layer="324" to-port="1" />
<edge from-layer="32" from-port="0" to-layer="88" to-port="1" />
<edge from-layer="32" from-port="0" to-layer="271" to-port="1" />
<edge from-layer="32" from-port="0" to-layer="218" to-port="1" />
<edge from-layer="32" from-port="0" to-layer="34" to-port="1" />
<edge from-layer="33" from-port="0" to-layer="218" to-port="2" />
<edge from-layer="33" from-port="0" to-layer="165" to-port="2" />
<edge from-layer="33" from-port="0" to-layer="271" to-port="2" />
<edge from-layer="33" from-port="0" to-layer="324" to-port="2" />
<edge from-layer="33" from-port="0" to-layer="34" to-port="2" />
<edge from-layer="34" from-port="5" to-layer="63" to-port="0" />
<edge from-layer="34" from-port="3" to-layer="36" to-port="0" />
<edge from-layer="34" from-port="4" to-layer="49" to-port="0" />
<edge from-layer="35" from-port="0" to-layer="36" to-port="1" />
<edge from-layer="36" from-port="2" to-layer="38" to-port="0" />
<edge from-layer="37" from-port="0" to-layer="38" to-port="1" />
<edge from-layer="38" from-port="2" to-layer="145" to-port="0" />
<edge from-layer="39" from-port="0" to-layer="44" to-port="0" />
<edge from-layer="40" from-port="0" to-layer="43" to-port="1" />
<edge from-layer="41" from-port="0" to-layer="43" to-port="2" />
<edge from-layer="42" from-port="0" to-layer="43" to-port="3" />
<edge from-layer="43" from-port="4" to-layer="44" to-port="1" />
<edge from-layer="44" from-port="2" to-layer="45" to-port="0" />
<edge from-layer="45" from-port="1" to-layer="47" to-port="0" />
<edge from-layer="46" from-port="0" to-layer="47" to-port="2" />
<edge from-layer="47" from-port="3" to-layer="77" to-port="0" />
<edge from-layer="47" from-port="3" to-layer="52" to-port="0" />
<edge from-layer="48" from-port="0" to-layer="49" to-port="1" />
<edge from-layer="49" from-port="2" to-layer="51" to-port="0" />
<edge from-layer="50" from-port="0" to-layer="51" to-port="1" />
<edge from-layer="51" from-port="2" to-layer="52" to-port="1" />
<edge from-layer="52" from-port="2" to-layer="145" to-port="1" />
<edge from-layer="52" from-port="2" to-layer="353" to-port="0" />
<edge from-layer="53" from-port="0" to-layer="58" to-port="0" />
<edge from-layer="54" from-port="0" to-layer="57" to-port="1" />
<edge from-layer="55" from-port="0" to-layer="57" to-port="2" />
<edge from-layer="56" from-port="0" to-layer="57" to-port="3" />
<edge from-layer="57" from-port="4" to-layer="58" to-port="1" />
<edge from-layer="58" from-port="2" to-layer="59" to-port="0" />
<edge from-layer="59" from-port="1" to-layer="61" to-port="0" />
<edge from-layer="60" from-port="0" to-layer="61" to-port="2" />
<edge from-layer="61" from-port="3" to-layer="66" to-port="0" />
<edge from-layer="62" from-port="0" to-layer="63" to-port="1" />
<edge from-layer="63" from-port="2" to-layer="65" to-port="0" />
<edge from-layer="64" from-port="0" to-layer="65" to-port="1" />
<edge from-layer="65" from-port="2" to-layer="66" to-port="1" />
<edge from-layer="66" from-port="2" to-layer="352" to-port="0" />
<edge from-layer="66" from-port="2" to-layer="145" to-port="2" />
<edge from-layer="67" from-port="0" to-layer="98" to-port="0" />
<edge from-layer="68" from-port="0" to-layer="83" to-port="0" />
<edge from-layer="69" from-port="1" to-layer="112" to-port="0" />
<edge from-layer="69" from-port="1" to-layer="72" to-port="0" />
<edge from-layer="70" from-port="0" to-layer="72" to-port="1" />
<edge from-layer="71" from-port="0" to-layer="72" to-port="2" />
<edge from-layer="72" from-port="3" to-layer="90" to-port="1" />
<edge from-layer="72" from-port="3" to-layer="74" to-port="0" />
<edge from-layer="73" from-port="0" to-layer="74" to-port="1" />
<edge from-layer="74" from-port="2" to-layer="133" to-port="2" />
<edge from-layer="74" from-port="2" to-layer="76" to-port="0" />
<edge from-layer="74" from-port="2" to-layer="141" to-port="1" />
<edge from-layer="75" from-port="0" to-layer="76" to-port="1" />
<edge from-layer="76" from-port="2" to-layer="81" to-port="0" />
<edge from-layer="77" from-port="1" to-layer="80" to-port="0" />
<edge from-layer="78" from-port="0" to-layer="80" to-port="1" />
<edge from-layer="79" from-port="0" to-layer="80" to-port="2" />
<edge from-layer="80" from-port="3" to-layer="90" to-port="0" />
<edge from-layer="80" from-port="3" to-layer="92" to-port="0" />
<edge from-layer="80" from-port="3" to-layer="140" to-port="0" />
<edge from-layer="80" from-port="3" to-layer="81" to-port="1" />
<edge from-layer="81" from-port="2" to-layer="83" to-port="1" />
<edge from-layer="81" from-port="2" to-layer="132" to-port="0" />
<edge from-layer="82" from-port="0" to-layer="83" to-port="2" />
<edge from-layer="83" from-port="3" to-layer="85" to-port="0" />
<edge from-layer="84" from-port="0" to-layer="93" to-port="1" />
<edge from-layer="84" from-port="0" to-layer="85" to-port="1" />
<edge from-layer="85" from-port="2" to-layer="87" to-port="0" />
<edge from-layer="86" from-port="0" to-layer="87" to-port="1" />
<edge from-layer="86" from-port="0" to-layer="116" to-port="1" />
<edge from-layer="86" from-port="0" to-layer="94" to-port="1" />
<edge from-layer="87" from-port="2" to-layer="88" to-port="0" />
<edge from-layer="88" from-port="2" to-layer="89" to-port="0" />
<edge from-layer="88" from-port="2" to-layer="108" to-port="0" />
<edge from-layer="89" from-port="1" to-layer="97" to-port="0" />
<edge from-layer="90" from-port="2" to-layer="92" to-port="1" />
<edge from-layer="91" from-port="0" to-layer="92" to-port="2" />
<edge from-layer="92" from-port="3" to-layer="93" to-port="0" />
<edge from-layer="93" from-port="2" to-layer="94" to-port="0" />
<edge from-layer="94" from-port="2" to-layer="96" to-port="0" />
<edge from-layer="95" from-port="0" to-layer="118" to-port="1" />
<edge from-layer="95" from-port="0" to-layer="96" to-port="1" />
<edge from-layer="96" from-port="2" to-layer="97" to-port="1" />
<edge from-layer="97" from-port="2" to-layer="98" to-port="1" />
<edge from-layer="98" from-port="2" to-layer="129" to-port="0" />
<edge from-layer="99" from-port="0" to-layer="100" to-port="1" />
<edge from-layer="100" from-port="2" to-layer="101" to-port="0" />
<edge from-layer="101" from-port="1" to-layer="102" to-port="0" />
<edge from-layer="101" from-port="1" to-layer="107" to-port="0" />
<edge from-layer="102" from-port="1" to-layer="121" to-port="0" />
<edge from-layer="102" from-port="1" to-layer="104" to-port="0" />
<edge from-layer="103" from-port="0" to-layer="104" to-port="1" />
<edge from-layer="104" from-port="2" to-layer="106" to-port="0" />
<edge from-layer="105" from-port="0" to-layer="106" to-port="1" />
<edge from-layer="106" from-port="2" to-layer="107" to-port="1" />
<edge from-layer="107" from-port="2" to-layer="124" to-port="0" />
<edge from-layer="108" from-port="1" to-layer="123" to-port="0" />
<edge from-layer="109" from-port="0" to-layer="115" to-port="0" />
<edge from-layer="110" from-port="0" to-layer="112" to-port="1" />
<edge from-layer="111" from-port="0" to-layer="112" to-port="2" />
<edge from-layer="112" from-port="3" to-layer="133" to-port="0" />
<edge from-layer="112" from-port="3" to-layer="113" to-port="0" />
<edge from-layer="113" from-port="1" to-layer="115" to-port="1" />
<edge from-layer="114" from-port="0" to-layer="115" to-port="2" />
<edge from-layer="115" from-port="3" to-layer="116" to-port="0" />
<edge from-layer="116" from-port="2" to-layer="117" to-port="0" />
<edge from-layer="117" from-port="2" to-layer="118" to-port="0" />
<edge from-layer="118" from-port="2" to-layer="119" to-port="0" />
<edge from-layer="119" from-port="1" to-layer="122" to-port="0" />
<edge from-layer="120" from-port="0" to-layer="124" to-port="2" />
<edge from-layer="120" from-port="0" to-layer="121" to-port="1" />
<edge from-layer="121" from-port="3" to-layer="122" to-port="1" />
<edge from-layer="122" from-port="2" to-layer="123" to-port="1" />
<edge from-layer="123" from-port="2" to-layer="127" to-port="0" />
<edge from-layer="123" from-port="2" to-layer="124" to-port="1" />
<edge from-layer="124" from-port="3" to-layer="126" to-port="0" />
<edge from-layer="125" from-port="0" to-layer="126" to-port="1" />
<edge from-layer="126" from-port="2" to-layer="128" to-port="0" />
<edge from-layer="127" from-port="1" to-layer="128" to-port="1" />
<edge from-layer="128" from-port="2" to-layer="129" to-port="1" />
<edge from-layer="129" from-port="2" to-layer="134" to-port="0" />
<edge from-layer="130" from-port="0" to-layer="133" to-port="1" />
<edge from-layer="131" from-port="0" to-layer="132" to-port="1" />
<edge from-layer="132" from-port="2" to-layer="133" to-port="3" />
<edge from-layer="133" from-port="4" to-layer="134" to-port="1" />
<edge from-layer="134" from-port="2" to-layer="137" to-port="0" />
<edge from-layer="135" from-port="0" to-layer="137" to-port="1" />
<edge from-layer="136" from-port="0" to-layer="137" to-port="2" />
<edge from-layer="137" from-port="3" to-layer="144" to-port="0" />
<edge from-layer="138" from-port="0" to-layer="144" to-port="1" />
<edge from-layer="139" from-port="0" to-layer="140" to-port="1" />
<edge from-layer="140" from-port="2" to-layer="141" to-port="0" />
<edge from-layer="141" from-port="2" to-layer="144" to-port="2" />
<edge from-layer="142" from-port="0" to-layer="144" to-port="3" />
<edge from-layer="143" from-port="0" to-layer="144" to-port="4" />
<edge from-layer="144" from-port="5" to-layer="304" to-port="3" />
<edge from-layer="144" from-port="5" to-layer="358" to-port="3" />
<edge from-layer="144" from-port="5" to-layer="198" to-port="3" />
<edge from-layer="144" from-port="5" to-layer="145" to-port="3" />
<edge from-layer="144" from-port="5" to-layer="251" to-port="3" />
<edge from-layer="145" from-port="4" to-layer="147" to-port="0" />
<edge from-layer="146" from-port="0" to-layer="147" to-port="1" />
<edge from-layer="147" from-port="2" to-layer="149" to-port="0" />
<edge from-layer="148" from-port="0" to-layer="149" to-port="1" />
<edge from-layer="149" from-port="2" to-layer="151" to-port="0" />
<edge from-layer="150" from-port="0" to-layer="151" to-port="1" />
<edge from-layer="151" from-port="2" to-layer="152" to-port="0" />
<edge from-layer="152" from-port="2" to-layer="160" to-port="0" />
<edge from-layer="152" from-port="2" to-layer="154" to-port="0" />
<edge from-layer="153" from-port="0" to-layer="154" to-port="1" />
<edge from-layer="154" from-port="2" to-layer="156" to-port="0" />
<edge from-layer="155" from-port="0" to-layer="156" to-port="1" />
<edge from-layer="156" from-port="2" to-layer="157" to-port="0" />
<edge from-layer="157" from-port="1" to-layer="159" to-port="0" />
<edge from-layer="158" from-port="0" to-layer="159" to-port="1" />
<edge from-layer="159" from-port="2" to-layer="160" to-port="1" />
<edge from-layer="160" from-port="2" to-layer="162" to-port="0" />
<edge from-layer="160" from-port="2" to-layer="205" to-port="1" />
<edge from-layer="161" from-port="0" to-layer="162" to-port="1" />
<edge from-layer="162" from-port="2" to-layer="164" to-port="0" />
<edge from-layer="163" from-port="0" to-layer="164" to-port="1" />
<edge from-layer="164" from-port="2" to-layer="165" to-port="0" />
<edge from-layer="165" from-port="5" to-layer="194" to-port="0" />
<edge from-layer="165" from-port="3" to-layer="167" to-port="0" />
<edge from-layer="165" from-port="4" to-layer="180" to-port="0" />
<edge from-layer="166" from-port="0" to-layer="167" to-port="1" />
<edge from-layer="167" from-port="2" to-layer="169" to-port="0" />
<edge from-layer="168" from-port="0" to-layer="169" to-port="1" />
<edge from-layer="169" from-port="2" to-layer="198" to-port="0" />
<edge from-layer="170" from-port="0" to-layer="175" to-port="0" />
<edge from-layer="171" from-port="0" to-layer="174" to-port="1" />
<edge from-layer="172" from-port="0" to-layer="174" to-port="2" />
<edge from-layer="173" from-port="0" to-layer="174" to-port="3" />
<edge from-layer="174" from-port="4" to-layer="175" to-port="1" />
<edge from-layer="175" from-port="2" to-layer="176" to-port="0" />
<edge from-layer="176" from-port="1" to-layer="178" to-port="0" />
<edge from-layer="177" from-port="0" to-layer="178" to-port="2" />
<edge from-layer="178" from-port="3" to-layer="183" to-port="0" />
<edge from-layer="179" from-port="0" to-layer="180" to-port="1" />
<edge from-layer="180" from-port="2" to-layer="182" to-port="0" />
<edge from-layer="181" from-port="0" to-layer="182" to-port="1" />
<edge from-layer="182" from-port="2" to-layer="183" to-port="1" />
<edge from-layer="183" from-port="2" to-layer="351" to-port="0" />
<edge from-layer="183" from-port="2" to-layer="198" to-port="1" />
<edge from-layer="184" from-port="0" to-layer="189" to-port="0" />
<edge from-layer="185" from-port="0" to-layer="188" to-port="1" />
<edge from-layer="186" from-port="0" to-layer="188" to-port="2" />
<edge from-layer="187" from-port="0" to-layer="188" to-port="3" />
<edge from-layer="188" from-port="4" to-layer="189" to-port="1" />
<edge from-layer="189" from-port="2" to-layer="190" to-port="0" />
<edge from-layer="190" from-port="1" to-layer="192" to-port="0" />
<edge from-layer="191" from-port="0" to-layer="192" to-port="2" />
<edge from-layer="192" from-port="3" to-layer="197" to-port="0" />
<edge from-layer="193" from-port="0" to-layer="194" to-port="1" />
<edge from-layer="194" from-port="2" to-layer="196" to-port="0" />
<edge from-layer="195" from-port="0" to-layer="196" to-port="1" />
<edge from-layer="196" from-port="2" to-layer="197" to-port="1" />
<edge from-layer="197" from-port="2" to-layer="198" to-port="2" />
<edge from-layer="197" from-port="2" to-layer="350" to-port="0" />
<edge from-layer="198" from-port="4" to-layer="200" to-port="0" />
<edge from-layer="199" from-port="0" to-layer="200" to-port="1" />
<edge from-layer="200" from-port="2" to-layer="202" to-port="0" />
<edge from-layer="201" from-port="0" to-layer="202" to-port="1" />
<edge from-layer="202" from-port="2" to-layer="204" to-port="0" />
<edge from-layer="203" from-port="0" to-layer="204" to-port="1" />
<edge from-layer="204" from-port="2" to-layer="205" to-port="0" />
<edge from-layer="205" from-port="2" to-layer="213" to-port="0" />
<edge from-layer="205" from-port="2" to-layer="207" to-port="0" />
<edge from-layer="206" from-port="0" to-layer="207" to-port="1" />
<edge from-layer="207" from-port="2" to-layer="209" to-port="0" />
<edge from-layer="208" from-port="0" to-layer="209" to-port="1" />
<edge from-layer="209" from-port="2" to-layer="210" to-port="0" />
<edge from-layer="210" from-port="1" to-layer="212" to-port="0" />
<edge from-layer="211" from-port="0" to-layer="212" to-port="1" />
<edge from-layer="212" from-port="2" to-layer="213" to-port="1" />
<edge from-layer="213" from-port="2" to-layer="215" to-port="0" />
<edge from-layer="213" from-port="2" to-layer="258" to-port="1" />
<edge from-layer="214" from-port="0" to-layer="215" to-port="1" />
<edge from-layer="215" from-port="2" to-layer="217" to-port="0" />
<edge from-layer="216" from-port="0" to-layer="217" to-port="1" />
<edge from-layer="217" from-port="2" to-layer="218" to-port="0" />
<edge from-layer="218" from-port="3" to-layer="220" to-port="0" />
<edge from-layer="218" from-port="4" to-layer="233" to-port="0" />
<edge from-layer="218" from-port="5" to-layer="247" to-port="0" />
<edge from-layer="219" from-port="0" to-layer="220" to-port="1" />
<edge from-layer="220" from-port="2" to-layer="222" to-port="0" />
<edge from-layer="221" from-port="0" to-layer="222" to-port="1" />
<edge from-layer="222" from-port="2" to-layer="251" to-port="0" />
<edge from-layer="223" from-port="0" to-layer="228" to-port="0" />
<edge from-layer="224" from-port="0" to-layer="227" to-port="1" />
<edge from-layer="225" from-port="0" to-layer="227" to-port="2" />
<edge from-layer="226" from-port="0" to-layer="227" to-port="3" />
<edge from-layer="227" from-port="4" to-layer="228" to-port="1" />
<edge from-layer="228" from-port="2" to-layer="229" to-port="0" />
<edge from-layer="229" from-port="1" to-layer="231" to-port="0" />
<edge from-layer="230" from-port="0" to-layer="231" to-port="2" />
<edge from-layer="231" from-port="3" to-layer="236" to-port="0" />
<edge from-layer="232" from-port="0" to-layer="233" to-port="1" />
<edge from-layer="233" from-port="2" to-layer="235" to-port="0" />
<edge from-layer="234" from-port="0" to-layer="235" to-port="1" />
<edge from-layer="235" from-port="2" to-layer="236" to-port="1" />
<edge from-layer="236" from-port="2" to-layer="251" to-port="1" />
<edge from-layer="236" from-port="2" to-layer="349" to-port="0" />
<edge from-layer="237" from-port="0" to-layer="242" to-port="0" />
<edge from-layer="238" from-port="0" to-layer="241" to-port="1" />
<edge from-layer="239" from-port="0" to-layer="241" to-port="2" />
<edge from-layer="240" from-port="0" to-layer="241" to-port="3" />
<edge from-layer="241" from-port="4" to-layer="242" to-port="1" />
<edge from-layer="242" from-port="2" to-layer="243" to-port="0" />
<edge from-layer="243" from-port="1" to-layer="245" to-port="0" />
<edge from-layer="244" from-port="0" to-layer="245" to-port="2" />
<edge from-layer="245" from-port="3" to-layer="250" to-port="0" />
<edge from-layer="246" from-port="0" to-layer="247" to-port="1" />
<edge from-layer="247" from-port="2" to-layer="249" to-port="0" />
<edge from-layer="248" from-port="0" to-layer="249" to-port="1" />
<edge from-layer="249" from-port="2" to-layer="250" to-port="1" />
<edge from-layer="250" from-port="2" to-layer="251" to-port="2" />
<edge from-layer="250" from-port="2" to-layer="348" to-port="0" />
<edge from-layer="251" from-port="4" to-layer="253" to-port="0" />
<edge from-layer="252" from-port="0" to-layer="253" to-port="1" />
<edge from-layer="253" from-port="2" to-layer="255" to-port="0" />
<edge from-layer="254" from-port="0" to-layer="255" to-port="1" />
<edge from-layer="255" from-port="2" to-layer="257" to-port="0" />
<edge from-layer="256" from-port="0" to-layer="257" to-port="1" />
<edge from-layer="257" from-port="2" to-layer="258" to-port="0" />
<edge from-layer="258" from-port="2" to-layer="260" to-port="0" />
<edge from-layer="258" from-port="2" to-layer="266" to-port="0" />
<edge from-layer="259" from-port="0" to-layer="260" to-port="1" />
<edge from-layer="260" from-port="2" to-layer="262" to-port="0" />
<edge from-layer="261" from-port="0" to-layer="262" to-port="1" />
<edge from-layer="262" from-port="2" to-layer="263" to-port="0" />
<edge from-layer="263" from-port="1" to-layer="265" to-port="0" />
<edge from-layer="264" from-port="0" to-layer="265" to-port="1" />
<edge from-layer="265" from-port="2" to-layer="266" to-port="1" />
<edge from-layer="266" from-port="2" to-layer="268" to-port="0" />
<edge from-layer="266" from-port="2" to-layer="311" to-port="1" />
<edge from-layer="267" from-port="0" to-layer="268" to-port="1" />
<edge from-layer="268" from-port="2" to-layer="270" to-port="0" />
<edge from-layer="269" from-port="0" to-layer="270" to-port="1" />
<edge from-layer="270" from-port="2" to-layer="271" to-port="0" />
<edge from-layer="271" from-port="4" to-layer="286" to-port="0" />
<edge from-layer="271" from-port="5" to-layer="300" to-port="0" />
<edge from-layer="271" from-port="3" to-layer="273" to-port="0" />
<edge from-layer="272" from-port="0" to-layer="273" to-port="1" />
<edge from-layer="273" from-port="2" to-layer="275" to-port="0" />
<edge from-layer="274" from-port="0" to-layer="275" to-port="1" />
<edge from-layer="275" from-port="2" to-layer="304" to-port="0" />
<edge from-layer="276" from-port="0" to-layer="281" to-port="0" />
<edge from-layer="277" from-port="0" to-layer="280" to-port="1" />
<edge from-layer="278" from-port="0" to-layer="280" to-port="2" />
<edge from-layer="279" from-port="0" to-layer="280" to-port="3" />
<edge from-layer="280" from-port="4" to-layer="281" to-port="1" />
<edge from-layer="281" from-port="2" to-layer="282" to-port="0" />
<edge from-layer="282" from-port="1" to-layer="284" to-port="0" />
<edge from-layer="283" from-port="0" to-layer="284" to-port="2" />
<edge from-layer="284" from-port="3" to-layer="289" to-port="0" />
<edge from-layer="285" from-port="0" to-layer="286" to-port="1" />
<edge from-layer="286" from-port="2" to-layer="288" to-port="0" />
<edge from-layer="287" from-port="0" to-layer="288" to-port="1" />
<edge from-layer="288" from-port="2" to-layer="289" to-port="1" />
<edge from-layer="289" from-port="2" to-layer="304" to-port="1" />
<edge from-layer="289" from-port="2" to-layer="347" to-port="0" />
<edge from-layer="290" from-port="0" to-layer="295" to-port="0" />
<edge from-layer="291" from-port="0" to-layer="294" to-port="1" />
<edge from-layer="292" from-port="0" to-layer="294" to-port="2" />
<edge from-layer="293" from-port="0" to-layer="294" to-port="3" />
<edge from-layer="294" from-port="4" to-layer="295" to-port="1" />
<edge from-layer="295" from-port="2" to-layer="296" to-port="0" />
<edge from-layer="296" from-port="1" to-layer="298" to-port="0" />
<edge from-layer="297" from-port="0" to-layer="298" to-port="2" />
<edge from-layer="298" from-port="3" to-layer="303" to-port="0" />
<edge from-layer="299" from-port="0" to-layer="300" to-port="1" />
<edge from-layer="300" from-port="2" to-layer="302" to-port="0" />
<edge from-layer="301" from-port="0" to-layer="302" to-port="1" />
<edge from-layer="302" from-port="2" to-layer="303" to-port="1" />
<edge from-layer="303" from-port="2" to-layer="304" to-port="2" />
<edge from-layer="303" from-port="2" to-layer="346" to-port="0" />
<edge from-layer="304" from-port="4" to-layer="306" to-port="0" />
<edge from-layer="305" from-port="0" to-layer="306" to-port="1" />
<edge from-layer="306" from-port="2" to-layer="308" to-port="0" />
<edge from-layer="307" from-port="0" to-layer="308" to-port="1" />
<edge from-layer="308" from-port="2" to-layer="310" to-port="0" />
<edge from-layer="309" from-port="0" to-layer="310" to-port="1" />
<edge from-layer="310" from-port="2" to-layer="311" to-port="0" />
<edge from-layer="311" from-port="2" to-layer="313" to-port="0" />
<edge from-layer="311" from-port="2" to-layer="319" to-port="0" />
<edge from-layer="312" from-port="0" to-layer="313" to-port="1" />
<edge from-layer="313" from-port="2" to-layer="315" to-port="0" />
<edge from-layer="314" from-port="0" to-layer="315" to-port="1" />
<edge from-layer="315" from-port="2" to-layer="316" to-port="0" />
<edge from-layer="316" from-port="1" to-layer="318" to-port="0" />
<edge from-layer="317" from-port="0" to-layer="318" to-port="1" />
<edge from-layer="318" from-port="2" to-layer="319" to-port="1" />
<edge from-layer="319" from-port="2" to-layer="365" to-port="1" />
<edge from-layer="319" from-port="2" to-layer="321" to-port="0" />
<edge from-layer="320" from-port="0" to-layer="321" to-port="1" />
<edge from-layer="321" from-port="2" to-layer="323" to-port="0" />
<edge from-layer="322" from-port="0" to-layer="323" to-port="1" />
<edge from-layer="323" from-port="2" to-layer="324" to-port="0" />
<edge from-layer="324" from-port="3" to-layer="355" to-port="0" />
<edge from-layer="324" from-port="5" to-layer="326" to-port="0" />
<edge from-layer="324" from-port="4" to-layer="341" to-port="0" />
<edge from-layer="325" from-port="0" to-layer="326" to-port="1" />
<edge from-layer="326" from-port="2" to-layer="328" to-port="0" />
<edge from-layer="327" from-port="0" to-layer="328" to-port="1" />
<edge from-layer="328" from-port="2" to-layer="329" to-port="1" />
<edge from-layer="329" from-port="2" to-layer="358" to-port="2" />
<edge from-layer="329" from-port="2" to-layer="330" to-port="0" />
<edge from-layer="331" from-port="0" to-layer="336" to-port="0" />
<edge from-layer="332" from-port="0" to-layer="335" to-port="1" />
<edge from-layer="333" from-port="0" to-layer="335" to-port="2" />
<edge from-layer="334" from-port="0" to-layer="335" to-port="3" />
<edge from-layer="335" from-port="4" to-layer="336" to-port="1" />
<edge from-layer="336" from-port="2" to-layer="337" to-port="0" />
<edge from-layer="337" from-port="1" to-layer="339" to-port="0" />
<edge from-layer="338" from-port="0" to-layer="339" to-port="2" />
<edge from-layer="339" from-port="3" to-layer="344" to-port="0" />
<edge from-layer="340" from-port="0" to-layer="341" to-port="1" />
<edge from-layer="341" from-port="2" to-layer="343" to-port="0" />
<edge from-layer="342" from-port="0" to-layer="343" to-port="1" />
<edge from-layer="343" from-port="2" to-layer="344" to-port="1" />
<edge from-layer="344" from-port="2" to-layer="345" to-port="0" />
<edge from-layer="344" from-port="2" to-layer="358" to-port="1" />
<edge from-layer="354" from-port="0" to-layer="355" to-port="1" />
<edge from-layer="355" from-port="2" to-layer="357" to-port="0" />
<edge from-layer="356" from-port="0" to-layer="357" to-port="1" />
<edge from-layer="357" from-port="2" to-layer="358" to-port="0" />
<edge from-layer="358" from-port="4" to-layer="360" to-port="0" />
<edge from-layer="359" from-port="0" to-layer="360" to-port="1" />
<edge from-layer="360" from-port="2" to-layer="362" to-port="0" />
<edge from-layer="361" from-port="0" to-layer="362" to-port="1" />
<edge from-layer="362" from-port="2" to-layer="364" to-port="0" />
<edge from-layer="363" from-port="0" to-layer="364" to-port="1" />
<edge from-layer="364" from-port="2" to-layer="365" to-port="0" />
<edge from-layer="365" from-port="2" to-layer="367" to-port="0" />
<edge from-layer="365" from-port="2" to-layer="373" to-port="0" />
<edge from-layer="366" from-port="0" to-layer="367" to-port="1" />
<edge from-layer="367" from-port="2" to-layer="369" to-port="0" />
<edge from-layer="368" from-port="0" to-layer="369" to-port="1" />
<edge from-layer="369" from-port="2" to-layer="370" to-port="0" />
<edge from-layer="370" from-port="1" to-layer="372" to-port="0" />
<edge from-layer="371" from-port="0" to-layer="372" to-port="1" />
<edge from-layer="372" from-port="2" to-layer="373" to-port="1" />
<edge from-layer="373" from-port="2" to-layer="375" to-port="0" />
<edge from-layer="374" from-port="0" to-layer="375" to-port="1" />
<edge from-layer="375" from-port="2" to-layer="377" to-port="0" />
<edge from-layer="376" from-port="0" to-layer="377" to-port="1" />
<edge from-layer="377" from-port="2" to-layer="378" to-port="0" />
<edge from-layer="378" from-port="2" to-layer="379" to-port="0" />
</edges>
<rt_info>
<Runtime_version value="2025.3.0-19807-44526285f24-releases/2025/3" />
<conversion_parameters>
<framework value="pytorch" />
<is_python_object value="True" />
</conversion_parameters>
<optimum>
<optimum_intel_version value="1.27.0.dev0+906008d" />
<optimum_version value="2.0.0" />
<pytorch_version value="2.9.1" />
<transformers_version value="4.55.4" />
</optimum>
<runtime_options>
<ACTIVATIONS_SCALE_FACTOR value="8.0" />
<KV_CACHE_PRECISION value="f16" />
</runtime_options>
</rt_info>
</net>