Hugging Face's logo

optimum-intel-internal-testing
/
tiny-random-muse-glimmer

Safetensors
muse_glimmer
Model card Files
xet
 Community
2
tiny-random-muse-glimmer / openvino_language_model.xml
echarlaix's picture
echarlaix HF Staff
Add reference OpenVINO IRs
b7a193b
Raw History Blame
413 kB
<?xml version="1.0"?>
<net name="Model6" version="11">
<layers>
<layer id="0" 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="2" name="inputs_embeds" type="Parameter" version="opset1">
<data shape="?,?,64" element_type="f32" />
<output>
<port id="0" precision="FP32" names="inputs_embeds">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="3" 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_44105" 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_44091" type="ShapeOf" version="opset3">
<data output_type="i64" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="6" name="Constant_44093" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="7" name="Constant_44095" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="8" name="Gather_44096" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>3</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_44098" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="10" name="Constant_44100" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="11" name="Constant_44102" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="12" name="Concat_44103" 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_44106" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="14" name="ReadValue_43841" type="ReadValue" version="opset6">
<data variable_id="past_key_values.3.valuepresent.3.value" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.3.value">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="15" name="Constant_43577" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="16" name="Gather_43578" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="17" name="Constant_40921" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="18" name="__module.model.language_model.embed_tokens.embed_norm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="77">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="19" name="Constant_18180" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="20" name="__module.model.language_model.embed_tokens.embed_norm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="79">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="21" name="Constant_40922" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="22" name="__module.model.language_model.embed_tokens.embed_norm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="80,mean_squared.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="23" name="Constant_40923" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="24" name="__module.model.language_model.embed_tokens.embed_norm/aten::pow/Power_1" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="81">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="25" name="__module.model.language_model.embed_tokens.embed_norm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="267,82,83,inputs_embeds_1,normed_output.1,x.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="26" name="Constant_40926" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="27" name="Constant_40924" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="28" name="__module.model.language_model.layers.0.input_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="268">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="29" name="Constant_18698" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="30" name="__module.model.language_model.layers.0.input_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="270">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="31" name="Constant_40925" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="32" name="__module.model.language_model.layers.0.input_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="271">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="33" name="__module.model.language_model.layers.0.input_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="34" name="__module.model.language_model.layers.0.input_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="272">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="35" name="__module.model.language_model.layers.0.input_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="273,output.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="36" name="Constant_40927" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="37" name="__module.model.language_model.layers.0.input_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="276,277,hidden_states.3,output.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="38" name="self.model.language_model.layers.0.self_attn.q_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="316" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.self_attn.q_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="39" name="__module.model.language_model.layers.0.self_attn.q_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="40" name="__module.model.language_model.layers.0.self_attn.q_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="289">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="41" name="Constant_42095" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="8512" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="42" name="__module.model.language_model.layers.0.self_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>64</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="291">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="43" name="__module.model.language_model.layers.0.self_attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="44" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::to/Convert" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="303,hidden_states.7">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="45" name="Constant_40928" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="46" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="304">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="47" name="Constant_18905" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="48" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="306">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="49" name="Constant_40929" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="50" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="307,mean_squared.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="51" name="Constant_40930" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="52" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::pow/Power_1" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="308">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="53" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="309,310,normed_output.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="54" name="Constant_40931" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="8560" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="55" name="__module.model.language_model.layers.0.self_attn/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="311,q.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="56" name="__module.model.language_model.rotary_emb/aten::to/Convert" type="Const" version="opset1">
<data element_type="f32" shape="1, 8, 1" offset="8564" size="32" />
<output>
<port id="0" precision="FP32" names="236">
<dim>1</dim>
<dim>8</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="57" name="88" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8600" size="8" />
<output>
<port id="0" precision="I64" names="88" />
</output>
</layer>
<layer id="58" name="__module.model.language_model.rotary_emb/aten::unsqueeze/Unsqueeze_2" 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="242,243">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="59" name="__module.model.language_model.rotary_emb/aten::to/Convert_3" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="244,246,position_ids_expanded,position_ids_expanded.1">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="60" name="__module.model.language_model.rotary_emb/aten::matmul/MatMul" type="MatMul" version="opset1">
<data transpose_a="false" transpose_b="false" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>8</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="247">
<dim>-1</dim>
<dim>8</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="61" name="__module.model.language_model.rotary_emb/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="3" offset="8608" size="12" />
<output>
<port id="0" precision="I32">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="62" name="__module.model.language_model.rotary_emb/aten::transpose/Transpose" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>8</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="248,freqs">
<dim>-1</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="63" name="__module.model.language_model.rotary_emb/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-1" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="250,emb">
<dim>-1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="64" name="__module.model.language_model.rotary_emb/aten::cos/Cos" type="Cos" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="251,252,255,cos.1,cos.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="65" name="__module.model.language_model.layers.0.self_attn/aten::unsqueeze/Unsqueeze" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="320,522,724,cos,cos.5,cos.7">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="66" name="__module.model.language_model.layers.0.self_attn/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="322">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="67" name="Constant_19092" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="68" name="Constant_19094" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8632" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="69" name="Constant_19096" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="70" name="Constant_19048" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="71" name="__module.model.language_model.layers.0.self_attn/aten::slice/Slice" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="332,x2.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="72" name="__module.model.language_model.layers.0.self_attn/aten::neg/Constant" type="Const" version="opset1">
<data element_type="i32" shape="" offset="8648" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="73" name="__module.model.language_model.layers.0.self_attn/aten::neg/ConvertLike" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I32" />
</input>
<output>
<port id="1" precision="FP32" />
</output>
</layer>
<layer id="74" name="__module.model.language_model.layers.0.self_attn/aten::neg/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32" />
</input>
<output>
<port id="2" precision="FP32" names="333">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="75" name="Constant_18989" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="76" name="Constant_19033" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="77" name="Constant_19035" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="78" name="Constant_18987" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="79" name="__module.model.language_model.layers.0.self_attn/aten::slice/Slice_1" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="327,x1.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="80" name="__module.model.language_model.layers.0.self_attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-1" />
<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="335">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="81" name="__module.model.language_model.rotary_emb/aten::sin/Sin" type="Sin" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="253,254,256,sin.1,sin.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="82" name="__module.model.language_model.layers.0.self_attn/aten::unsqueeze/Unsqueeze_1" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="321,523,725,sin,sin.5,sin.7">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="83" name="__module.model.language_model.layers.0.self_attn/aten::mul/Multiply_2" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="336">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="84" name="__module.model.language_model.layers.0.self_attn/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="337,query.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="85" name="Constant_44117" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="86" name="Constant_44110" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="87" name="Constant_44112" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="88" name="Constant_44114" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="89" name="Concat_44115" 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="90" name="Broadcast_44118" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="91" name="ReadValue_43827" type="ReadValue" version="opset6">
<data variable_id="past_key_values.0.keypresent.0.key" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.0.key">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="92" name="Constant_43556" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="93" name="Gather_43557" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="94" name="self.model.language_model.layers.0.self_attn.k_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="8652" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.self_attn.k_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="95" name="__module.model.language_model.layers.0.self_attn.k_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="96" name="__module.model.language_model.layers.0.self_attn.k_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="294">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="97" name="Constant_42096" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="98" name="__module.model.language_model.layers.0.self_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="296">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="99" name="__module.model.language_model.layers.0.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="100" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::to/Convert_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="312,hidden_states.11">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="101" name="Constant_40932" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="102" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::pow/Power_2" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="313">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="103" name="Constant_18951" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="104" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::mean/ReduceMean_1" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="315">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="105" name="Constant_40933" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="106" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="316,mean_squared.5">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="107" name="Constant_40934" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="108" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::pow/Power_3" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="317">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="109" name="__module.model.language_model.layers.0.self_attn.qk_norm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="318,319,k.1,normed_output.5">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="110" name="__module.model.language_model.layers.0.self_attn/aten::mul/Multiply_3" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="338">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="111" name="Constant_19234" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="112" name="Constant_19236" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8632" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="113" name="Constant_19238" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="114" name="Constant_19190" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="115" name="__module.model.language_model.layers.0.self_attn/aten::slice/Slice_2" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="348,x2.3">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="116" name="__module.model.language_model.layers.0.self_attn/aten::neg/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="" offset="8648" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="117" name="__module.model.language_model.layers.0.self_attn/aten::neg/ConvertLike_1" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I32" />
</input>
<output>
<port id="1" precision="FP32" />
</output>
</layer>
<layer id="118" name="__module.model.language_model.layers.0.self_attn/aten::neg/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32" />
</input>
<output>
<port id="2" precision="FP32" names="349">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="119" name="Constant_19131" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="120" name="Constant_19175" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="121" name="Constant_19177" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="122" name="Constant_19129" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="123" name="__module.model.language_model.layers.0.self_attn/aten::slice/Slice_3" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="343,x1.3">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="124" name="__module.model.language_model.layers.0.self_attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-1" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="351">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="125" name="__module.model.language_model.layers.0.self_attn/aten::mul/Multiply_4" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="352">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="126" name="__module.model.language_model.layers.0.self_attn/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="353,key_states.1">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="127" name="__module.model.language_model.layers.0.self_attn/aten::cat/Concat_2" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="363,364,hidden_states.13">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="128" name="89" type="Const" version="opset1">
<data element_type="i64" shape="" offset="16" size="8" />
<output>
<port id="0" precision="I64" names="89" />
</output>
</layer>
<layer id="129" name="__module.model.language_model.layers.0.self_attn/aten::unsqueeze/Unsqueeze_2" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="365,366,367">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="130" name="__module.model.language_model.layers.0.self_attn/aten::size/ShapeOf_2" type="ShapeOf" version="opset3">
<data output_type="i64" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="131" name="Constant_38943" type="Const" version="opset1">
<data element_type="i64" shape="2" offset="12784" size="16" />
<output>
<port id="0" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="132" name="Constant_38944" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="133" name="Gather_38945" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>4</dim>
</port>
<port id="1" precision="I64">
<dim>2</dim>
</port>
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64">
<dim>2</dim>
</port>
</output>
</layer>
<layer id="134" name="Constant_38699" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="135" name="ShapeOf_41397" type="ShapeOf" version="opset3">
<data output_type="i64" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="136" name="Constant_41398" type="Const" version="opset1">
<data element_type="i64" shape="" offset="16" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="137" name="Constant_41399" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="138" name="Gather_41400" 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="125,127,129,176,178,180,q_offset.3,q_offset.9" />
</output>
</layer>
<layer id="139" name="Constant_41419" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="140" name="Reshape_41420" 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="ShapeOf_41403" type="ShapeOf" version="opset3">
<data output_type="i64" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="142" name="Constant_41404" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8600" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="143" name="Constant_41405" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="144" name="Gather_41406" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>3</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64" names="123,174" />
</output>
</layer>
<layer id="145" name="Constant_41421" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="146" name="Reshape_41422" 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="147" name="Add_41423" 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="361,563,765,933">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="148" name="Constant_38702" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="149" name="__module.model.language_model.layers.0.self_attn/prim::ListConstruct/SequenceMark_2" type="Concat" version="opset1">
<data axis="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">
<dim>1</dim>
</port>
<port id="3" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="4" precision="I64">
<dim>5</dim>
</port>
</output>
</layer>
<layer id="150" name="__module.model.language_model.layers.0.self_attn/aten::expand/Broadcast" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="369,hidden_states.15">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="151" name="Constant_42097" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="152" name="__module.model.language_model.layers.0.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="373,key.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="153" name="Constant_44129" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="154" name="Constant_44122" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="155" name="Constant_44124" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="156" name="Constant_44126" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="157" name="Concat_44127" 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="158" name="Broadcast_44130" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="159" name="ReadValue_43829" type="ReadValue" version="opset6">
<data variable_id="past_key_values.0.valuepresent.0.value" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.0.value">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="160" name="Constant_43559" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="161" name="Gather_43560" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="162" name="self.model.language_model.layers.0.self_attn.v_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="12832" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.self_attn.v_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="163" name="__module.model.language_model.layers.0.self_attn.v_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="164" name="__module.model.language_model.layers.0.self_attn.v_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="299">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="165" name="Constant_42098" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="166" name="__module.model.language_model.layers.0.self_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="301">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="167" name="__module.model.language_model.layers.0.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="168" name="__module.model.language_model.layers.0.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="302,value_states.1">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="169" name="__module.model.language_model.layers.0.self_attn/aten::cat/Concat_3" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="379,380,hidden_states.17">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="170" name="__module.model.language_model.layers.0.self_attn/aten::unsqueeze/Unsqueeze_3" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="381,382,383">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="171" name="__module.model.language_model.layers.0.self_attn/aten::expand/Broadcast_1" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="385,hidden_states.19">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="172" name="Constant_42099" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="173" name="__module.model.language_model.layers.0.self_attn/aten::reshape/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="389,value.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="174" name="__module.model.language_model/aten::new_ones/Broadcast" type="Const" version="opset1">
<data element_type="boolean" shape="" offset="16928" size="1" />
<output>
<port id="0" precision="BOOL" names="209,result.9" />
</output>
</layer>
<layer id="175" name="__module.model.language_model/aten::new_ones/Broadcast_1" type="Const" version="opset1">
<data element_type="boolean" shape="" offset="16928" size="1" />
<output>
<port id="0" precision="BOOL" names="211,result.5" />
</output>
</layer>
<layer id="176" name="__module.model.language_model/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="177" name="__module.model.language_model/aten::add/Add" 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="131,132,135,182,183,186,kv_length,kv_length.1" />
</output>
</layer>
<layer id="178" name="__module.model.language_model/aten::arange/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="" offset="16932" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="179" name="__module.model.language_model/aten::arange/Range" 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="143,144,194,195,kv_indices,kv_indices.1">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="180" name="91" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" names="91" />
</output>
</layer>
<layer id="181" name="__module.model.language_model/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="153,204">
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="182" name="__module.model.language_model/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="154,205">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="183" name="__module.model.language_model/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="155,156,206,207,kv_idx,kv_idx.1">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="184" name="__module.model.language_model/aten::arange/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="185" name="__module.model.language_model/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="186" name="__module.model.language_model/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="138,189" />
</output>
</layer>
<layer id="187" name="__module.model.language_model/aten::arange/Constant_4" type="Const" version="opset1">
<data element_type="i32" shape="" offset="16932" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="188" name="__module.model.language_model/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="141,192">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="189" name="__module.model.language_model/aten::arange/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="" offset="16932" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="190" name="__module.model.language_model/aten::arange/Range_1" type="Range" version="opset4">
<data output_type="i64" />
<input>
<port id="0" precision="I64" />
<port id="1" precision="I64" />
<port id="2" precision="I32" />
</input>
<output>
<port id="3" precision="I64" names="137,188,cache_position,cache_position.1">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="191" name="__module.model.language_model/aten::select/Gather" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64" names="139,190,q_offset,q_offset.5" />
</output>
</layer>
<layer id="192" name="__module.model.language_model/aten::add/Add_3" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="I64" names="142,193,q_indices,q_indices.1">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="193" name="__module.model.language_model/aten::unsqueeze/Unsqueeze_3" 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="149,200">
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="194" name="__module.model.language_model/aten::unsqueeze/Unsqueeze_4" 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="150,151,201,202">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="195" name="94" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8640" size="8" />
<output>
<port id="0" precision="I64" names="94" />
</output>
</layer>
<layer id="196" name="__module.model.language_model/aten::unsqueeze/Unsqueeze_5" 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="152,203,q_idx,q_idx.1">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="197" name="Constant_40935" type="Const" version="opset1">
<data element_type="i64" shape="1, 1, 1, 1" offset="16936" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="198" name="__module.model.language_model/aten::sub/Subtract" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="I64" names="212">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="199" name="__module.model.language_model/aten::gt/Greater" type="Greater" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL" names="213,214">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="200" name="__module.model.language_model/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="215,result.7">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="201" name="__module.model.language_model/aten::le/LessEqual" type="LessEqual" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="BOOL" names="159,160,216,217">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="202" name="__module.model.language_model/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="218,219">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="203" name="__module.model.language_model/aten::__and__/BitwiseAnd_2" 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="220,result">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="204" name="__module.model.language_model/aten::to/Convert_4" 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="122,173,attention_mask.3,attention_mask.7">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="205" name="__module.model.language_model/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="206" name="Constant_23870" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="207" name="__module.model.language_model/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="208" name="Constant_23872" type="Const" version="opset1">
<data element_type="i32" shape="1" offset="8648" size="4" />
<output>
<port id="0" precision="I32">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="209" name="__module.model.language_model/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="210" name="__module.model.language_model/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="211" name="__module.model.language_model/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="212" name="__module.model.language_model/aten::arange/Constant_5" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="213" name="Constant_41408" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="214" name="Constant_41409" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="215" name="Gather_41410" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="I64">
<dim>3</dim>
</port>
<port id="1" precision="I64" />
<port id="2" precision="I64" />
</input>
<output>
<port id="3" precision="I64" names="134,185" />
</output>
</layer>
<layer id="216" name="__module.model.language_model/aten::arange/Constant_6" type="Const" version="opset1">
<data element_type="i32" shape="" offset="16932" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="217" name="__module.model.language_model/aten::arange/Range_3" 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="140,145,191,196,batch_indices,batch_indices.1">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="218" name="__module.model.language_model/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="146,197">
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="219" name="__module.model.language_model/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="147,198">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="220" name="__module.model.language_model/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="148,199,batch_idx,batch_idx.1">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="221" name="__module.model.language_model/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="222" name="__module.model.language_model/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="223" name="__module.model.language_model/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="224" name="__module.model.language_model/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="225" name="__module.model.language_model/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="226" name="__module.model.language_model/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="227" name="Constant_33029" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="228" name="__module.model.language_model/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="229" name="__module.model.language_model/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="230" name="__module.model.language_model/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="163,164,222,223">
<dim>-1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="231" name="__module.model.language_model/aten::__and__/BitwiseAnd_3" 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="224,attention_mask.9">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="232" name="Constant_41569" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="233" name="Reshape_41570" 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="234" name="Constant_38730" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="235" name="Constant_18531" type="Const" version="opset1">
<data element_type="i32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="236" name="Unsqueeze_18535" 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="237" name="__module.model.language_model/prim::ListConstruct/SequenceMark" 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="238" name="__module.model.language_model/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="226,mask">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="239" name="__module.model.language_model/aten::to/Convert_6" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" names="168,227" />
</output>
</layer>
<layer id="240" name="__module.model.language_model/aten::to/Convert_7" type="Const" version="opset1">
<data element_type="f32" shape="" offset="16944" size="4" />
<output>
<port id="0" precision="FP32" names="170,229" />
</output>
</layer>
<layer id="241" name="__module.model.language_model/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="231,attention_mask.11">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="242" name="__module.model.language_model.layers.0.self_attn/aten::scaled_dot_product_attention/ConvertLike" type="Const" version="opset1">
<data element_type="f32" shape="" offset="16948" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="243" name="__module.model.language_model.layers.0.self_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>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="4" precision="FP32" />
</input>
<output>
<port id="5" precision="FP32" names="390,attn_output.1">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="244" name="__module.model.language_model.layers.0.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="245" name="__module.model.language_model.layers.0.self_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>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="391">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="246" name="Constant_42100" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="16952" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="247" name="__module.model.language_model.layers.0.self_attn/aten::reshape/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="394">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="248" name="self.model.language_model.layers.0.self_attn.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="16976" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.self_attn.gate_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="249" name="__module.model.language_model.layers.0.self_attn.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="250" name="__module.model.language_model.layers.0.self_attn.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="397">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="251" name="__module.model.language_model.layers.0.self_attn/aten::sigmoid/Sigmoid" type="Sigmoid" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="398">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="252" name="__module.model.language_model.layers.0.self_attn/aten::mul/Multiply_7" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="399">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="253" name="self.model.language_model.layers.0.self_attn.o_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="25168" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.self_attn.o_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="254" name="__module.model.language_model.layers.0.self_attn.o_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="255" name="__module.model.language_model.layers.0.self_attn.o_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="401,407,x.3,x.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="256" name="Constant_40938" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="257" name="Constant_40936" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="258" name="__module.model.language_model.layers.0.post_attention_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="408">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="259" name="Constant_19783" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="260" name="__module.model.language_model.layers.0.post_attention_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="410">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="261" name="Constant_40937" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="262" name="__module.model.language_model.layers.0.post_attention_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="411">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="263" name="__module.model.language_model.layers.0.post_attention_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="264" name="__module.model.language_model.layers.0.post_attention_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="412">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="265" name="__module.model.language_model.layers.0.post_attention_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="413,output.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="266" name="Constant_40939" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="267" name="__module.model.language_model.layers.0.post_attention_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="416,417,hidden_states.21,output.7">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="268" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="418,420,x.7,x.9">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="269" name="Constant_40942" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="270" name="Constant_40940" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="271" name="__module.model.language_model.layers.0.pre_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="421">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="272" name="Constant_19848" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="273" name="__module.model.language_model.layers.0.pre_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="423">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="274" name="Constant_40941" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="275" name="__module.model.language_model.layers.0.pre_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="424">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="276" name="__module.model.language_model.layers.0.pre_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="277" name="__module.model.language_model.layers.0.pre_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="425">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="278" name="__module.model.language_model.layers.0.pre_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="426,output.9">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="279" name="Constant_40943" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="280" name="__module.model.language_model.layers.0.pre_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="429,430,output.11">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="281" name="self.model.language_model.layers.0.mlp.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="33364" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.mlp.gate_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="282" name="__module.model.language_model.layers.0.mlp.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="283" name="__module.model.language_model.layers.0.mlp.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="438,input.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="284" name="__module.model.language_model.layers.0.mlp.act_fn/aten::silu/Swish" type="Swish" version="opset4">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="439">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="285" name="self.model.language_model.layers.0.mlp.up_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="66132" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.mlp.up_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="286" name="__module.model.language_model.layers.0.mlp.up_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="287" name="__module.model.language_model.layers.0.mlp.up_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="441">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="288" name="__module.model.language_model.layers.0.mlp/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="442">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="289" name="self.model.language_model.layers.0.mlp.down_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 256" offset="98900" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.0.mlp.down_proj._openvino_orig_weight">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="290" name="__module.model.language_model.layers.0.mlp.down_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="291" name="__module.model.language_model.layers.0.mlp.down_proj/ov_ext::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>256</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="444,446,x.11,x.13">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="292" name="Constant_40946" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="293" name="Constant_40944" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="294" name="__module.model.language_model.layers.0.post_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="447">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="295" name="Constant_19928" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="296" name="__module.model.language_model.layers.0.post_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="449">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="297" name="Constant_40945" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="298" name="__module.model.language_model.layers.0.post_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="450">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="299" name="__module.model.language_model.layers.0.post_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="300" name="__module.model.language_model.layers.0.post_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="451">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="301" name="__module.model.language_model.layers.0.post_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="452,output.13">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="302" name="Constant_40947" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="303" name="__module.model.language_model.layers.0.post_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="455,456,hidden_states.23,output.15">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="304" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="457,469,x.15,x.17">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="305" name="Constant_40950" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="306" name="Constant_40948" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="307" name="__module.model.language_model.layers.1.input_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="470">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="308" name="Constant_20000" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="309" name="__module.model.language_model.layers.1.input_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="472">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="310" name="Constant_40949" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="311" name="__module.model.language_model.layers.1.input_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="473">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="312" name="__module.model.language_model.layers.1.input_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="313" name="__module.model.language_model.layers.1.input_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="474">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="314" name="__module.model.language_model.layers.1.input_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="475,output.17">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="315" name="Constant_40951" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="316" name="__module.model.language_model.layers.1.input_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="478,479,hidden_states.25,output.19">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="317" name="self.model.language_model.layers.1.self_attn.q_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="131668" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.self_attn.q_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="318" name="__module.model.language_model.layers.1.self_attn.q_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="319" name="__module.model.language_model.layers.1.self_attn.q_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="491">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="320" name="Constant_42101" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="8512" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="321" name="__module.model.language_model.layers.1.self_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>64</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="493">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="322" name="__module.model.language_model.layers.1.self_attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="323" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::to/Convert" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="505,hidden_states.29">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="324" name="Constant_40952" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="325" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="506">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="326" name="Constant_20207" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="327" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="508">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="328" name="Constant_40953" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="329" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="509,mean_squared.7">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="330" name="Constant_40954" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="331" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::pow/Power_1" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="510">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="332" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="511,512,normed_output.7">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="333" name="Constant_40955" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="8560" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="334" name="__module.model.language_model.layers.1.self_attn/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="513,q.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="335" name="__module.model.language_model.layers.1.self_attn/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="524">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="336" name="Constant_20394" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="337" name="Constant_20396" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8632" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="338" name="Constant_20398" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="339" name="Constant_20350" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="340" name="__module.model.language_model.layers.1.self_attn/aten::slice/Slice" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="534,x2.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="341" name="__module.model.language_model.layers.1.self_attn/aten::neg/Constant" type="Const" version="opset1">
<data element_type="i32" shape="" offset="8648" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="342" name="__module.model.language_model.layers.1.self_attn/aten::neg/ConvertLike" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I32" />
</input>
<output>
<port id="1" precision="FP32" />
</output>
</layer>
<layer id="343" name="__module.model.language_model.layers.1.self_attn/aten::neg/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32" />
</input>
<output>
<port id="2" precision="FP32" names="535">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="344" name="Constant_20291" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="345" name="Constant_20335" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="346" name="Constant_20337" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="347" name="Constant_20289" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="348" name="__module.model.language_model.layers.1.self_attn/aten::slice/Slice_1" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="529,x1.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="349" name="__module.model.language_model.layers.1.self_attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-1" />
<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="537">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="350" name="__module.model.language_model.layers.1.self_attn/aten::mul/Multiply_2" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="538">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="351" name="__module.model.language_model.layers.1.self_attn/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="539,query.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="352" name="Constant_44153" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="353" name="Constant_44146" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="354" name="Constant_44148" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="355" name="Constant_44150" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="356" name="Concat_44151" 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="357" name="Broadcast_44154" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="358" name="ReadValue_43831" type="ReadValue" version="opset6">
<data variable_id="past_key_values.1.keypresent.1.key" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.1.key">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="359" name="Constant_43562" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="360" name="Gather_43563" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="361" name="self.model.language_model.layers.1.self_attn.k_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="139860" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.self_attn.k_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="362" name="__module.model.language_model.layers.1.self_attn.k_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="363" name="__module.model.language_model.layers.1.self_attn.k_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="496">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="364" name="Constant_42102" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="365" name="__module.model.language_model.layers.1.self_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="498">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="366" name="__module.model.language_model.layers.1.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="367" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::to/Convert_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="514,hidden_states.33">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="368" name="Constant_40956" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="369" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::pow/Power_2" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="515">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="370" name="Constant_20253" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="371" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::mean/ReduceMean_1" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="517">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="372" name="Constant_40957" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="373" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="518,mean_squared.9">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="374" name="Constant_40958" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="375" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::pow/Power_3" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="519">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="376" name="__module.model.language_model.layers.1.self_attn.qk_norm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="520,521,k.3,normed_output.9">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="377" name="__module.model.language_model.layers.1.self_attn/aten::mul/Multiply_3" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="540">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="378" name="Constant_20536" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="379" name="Constant_20538" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8632" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="380" name="Constant_20540" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="381" name="Constant_20492" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="382" name="__module.model.language_model.layers.1.self_attn/aten::slice/Slice_2" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="550,x2.7">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="383" name="__module.model.language_model.layers.1.self_attn/aten::neg/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="" offset="8648" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="384" name="__module.model.language_model.layers.1.self_attn/aten::neg/ConvertLike_1" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I32" />
</input>
<output>
<port id="1" precision="FP32" />
</output>
</layer>
<layer id="385" name="__module.model.language_model.layers.1.self_attn/aten::neg/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32" />
</input>
<output>
<port id="2" precision="FP32" names="551">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="386" name="Constant_20433" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="387" name="Constant_20477" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="388" name="Constant_20479" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="389" name="Constant_20431" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="390" name="__module.model.language_model.layers.1.self_attn/aten::slice/Slice_3" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="545,x1.7">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="391" name="__module.model.language_model.layers.1.self_attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-1" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="553">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="392" name="__module.model.language_model.layers.1.self_attn/aten::mul/Multiply_4" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="554">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="393" name="__module.model.language_model.layers.1.self_attn/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="555,key_states.3">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="394" name="__module.model.language_model.layers.1.self_attn/aten::cat/Concat_2" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="565,566,hidden_states.35">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="395" name="__module.model.language_model.layers.1.self_attn/aten::unsqueeze/Unsqueeze_2" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="567,568,569">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="396" name="__module.model.language_model.layers.1.self_attn/aten::expand/Broadcast" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="571,hidden_states.37">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="397" name="Constant_42103" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="398" name="__module.model.language_model.layers.1.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="575,key.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="399" name="Constant_44141" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="400" name="Constant_44134" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="401" name="Constant_44136" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="402" name="Constant_44138" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="403" name="Concat_44139" 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="404" name="Broadcast_44142" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="405" name="ReadValue_43833" type="ReadValue" version="opset6">
<data variable_id="past_key_values.1.valuepresent.1.value" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.1.value">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="406" name="Constant_43565" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="407" name="Gather_43566" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="408" name="self.model.language_model.layers.1.self_attn.v_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="143956" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.self_attn.v_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="409" name="__module.model.language_model.layers.1.self_attn.v_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="410" name="__module.model.language_model.layers.1.self_attn.v_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="501">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="411" name="Constant_42104" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="412" name="__module.model.language_model.layers.1.self_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="503">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="413" name="__module.model.language_model.layers.1.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="414" name="__module.model.language_model.layers.1.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="504,value_states.3">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="415" name="__module.model.language_model.layers.1.self_attn/aten::cat/Concat_3" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="581,582,hidden_states.39">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="416" name="__module.model.language_model.layers.1.self_attn/aten::unsqueeze/Unsqueeze_3" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="583,584,585">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="417" name="__module.model.language_model.layers.1.self_attn/aten::expand/Broadcast_1" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="587,hidden_states.41">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="418" name="Constant_42105" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="419" name="__module.model.language_model.layers.1.self_attn/aten::reshape/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="591,value.3">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="420" name="__module.model.language_model.layers.1.self_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>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="4" precision="FP32" />
</input>
<output>
<port id="5" precision="FP32" names="592,attn_output.7">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="421" name="__module.model.language_model.layers.1.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="422" name="__module.model.language_model.layers.1.self_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>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="593">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="423" name="Constant_42106" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="16952" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="424" name="__module.model.language_model.layers.1.self_attn/aten::reshape/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="596">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="425" name="self.model.language_model.layers.1.self_attn.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="148052" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.self_attn.gate_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="426" name="__module.model.language_model.layers.1.self_attn.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="427" name="__module.model.language_model.layers.1.self_attn.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="599">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="428" name="__module.model.language_model.layers.1.self_attn/aten::sigmoid/Sigmoid" type="Sigmoid" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="600">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="429" name="__module.model.language_model.layers.1.self_attn/aten::mul/Multiply_7" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="601">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="430" name="self.model.language_model.layers.1.self_attn.o_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="156244" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.self_attn.o_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="431" name="__module.model.language_model.layers.1.self_attn.o_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="432" name="__module.model.language_model.layers.1.self_attn.o_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="603,609,x.19,x.21">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="433" name="Constant_40961" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="434" name="Constant_40959" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="435" name="__module.model.language_model.layers.1.post_attention_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="610">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="436" name="Constant_21085" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="437" name="__module.model.language_model.layers.1.post_attention_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="612">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="438" name="Constant_40960" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="439" name="__module.model.language_model.layers.1.post_attention_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="613">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="440" name="__module.model.language_model.layers.1.post_attention_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="441" name="__module.model.language_model.layers.1.post_attention_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="614">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="442" name="__module.model.language_model.layers.1.post_attention_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="615,output.21">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="443" name="Constant_40962" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="444" name="__module.model.language_model.layers.1.post_attention_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="618,619,hidden_states.43,output.23">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="445" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="620,622,x.23,x.25">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="446" name="Constant_40965" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="447" name="Constant_40963" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="448" name="__module.model.language_model.layers.1.pre_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="623">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="449" name="Constant_21150" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="450" name="__module.model.language_model.layers.1.pre_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="625">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="451" name="Constant_40964" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="452" name="__module.model.language_model.layers.1.pre_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="626">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="453" name="__module.model.language_model.layers.1.pre_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="454" name="__module.model.language_model.layers.1.pre_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="627">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="455" name="__module.model.language_model.layers.1.pre_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="628,output.25">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="456" name="Constant_40966" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="457" name="__module.model.language_model.layers.1.pre_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="631,632,output.27">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="458" name="self.model.language_model.layers.1.mlp.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="164436" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.mlp.gate_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="459" name="__module.model.language_model.layers.1.mlp.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="460" name="__module.model.language_model.layers.1.mlp.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="640,input.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="461" name="__module.model.language_model.layers.1.mlp.act_fn/aten::silu/Swish" type="Swish" version="opset4">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="641">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="462" name="self.model.language_model.layers.1.mlp.up_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="197204" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.mlp.up_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="463" name="__module.model.language_model.layers.1.mlp.up_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="464" name="__module.model.language_model.layers.1.mlp.up_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="643">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="465" name="__module.model.language_model.layers.1.mlp/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="644">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="466" name="self.model.language_model.layers.1.mlp.down_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 256" offset="229972" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.1.mlp.down_proj._openvino_orig_weight">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="467" name="__module.model.language_model.layers.1.mlp.down_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="468" name="__module.model.language_model.layers.1.mlp.down_proj/ov_ext::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>256</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="646,648,x.27,x.29">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="469" name="Constant_40969" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="470" name="Constant_40967" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="471" name="__module.model.language_model.layers.1.post_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="649">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="472" name="Constant_21230" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="473" name="__module.model.language_model.layers.1.post_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="651">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="474" name="Constant_40968" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="475" name="__module.model.language_model.layers.1.post_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="652">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="476" name="__module.model.language_model.layers.1.post_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="477" name="__module.model.language_model.layers.1.post_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="653">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="478" name="__module.model.language_model.layers.1.post_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="654,output.29">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="479" name="Constant_40970" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="480" name="__module.model.language_model.layers.1.post_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="657,658,hidden_states.45,output.31">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="481" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="659,671,x.31,x.33">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="482" name="Constant_40973" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="483" name="Constant_40971" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="484" name="__module.model.language_model.layers.2.input_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="672">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="485" name="Constant_21302" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="486" name="__module.model.language_model.layers.2.input_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="674">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="487" name="Constant_40972" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="488" name="__module.model.language_model.layers.2.input_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="675">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="489" name="__module.model.language_model.layers.2.input_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="490" name="__module.model.language_model.layers.2.input_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="676">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="491" name="__module.model.language_model.layers.2.input_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="677,output.33">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="492" name="Constant_40974" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="493" name="__module.model.language_model.layers.2.input_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="680,681,hidden_states.47,output.35">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="494" name="self.model.language_model.layers.2.self_attn.q_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="262740" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.self_attn.q_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="495" name="__module.model.language_model.layers.2.self_attn.q_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="496" name="__module.model.language_model.layers.2.self_attn.q_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="693">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="497" name="Constant_42107" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="8512" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="498" name="__module.model.language_model.layers.2.self_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>64</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="695">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="499" name="__module.model.language_model.layers.2.self_attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="500" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::to/Convert" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="707,hidden_states.51">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="501" name="Constant_40975" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="502" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="708">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="503" name="Constant_21509" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="504" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="710">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="505" name="Constant_40976" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="506" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="711,mean_squared.11">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="507" name="Constant_40977" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="508" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::pow/Power_1" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="712">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="509" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="713,714,normed_output.11">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="510" name="Constant_40978" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="8560" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="511" name="__module.model.language_model.layers.2.self_attn/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="715,q">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="512" name="__module.model.language_model.layers.2.self_attn/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="726">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="513" name="Constant_21696" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="514" name="Constant_21698" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8632" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="515" name="Constant_21700" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="516" name="Constant_21652" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="517" name="__module.model.language_model.layers.2.self_attn/aten::slice/Slice" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="736,x2.9">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="518" name="__module.model.language_model.layers.2.self_attn/aten::neg/Constant" type="Const" version="opset1">
<data element_type="i32" shape="" offset="8648" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="519" name="__module.model.language_model.layers.2.self_attn/aten::neg/ConvertLike" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I32" />
</input>
<output>
<port id="1" precision="FP32" />
</output>
</layer>
<layer id="520" name="__module.model.language_model.layers.2.self_attn/aten::neg/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32" />
</input>
<output>
<port id="2" precision="FP32" names="737">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="521" name="Constant_21593" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="522" name="Constant_21637" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="523" name="Constant_21639" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="524" name="Constant_21591" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="525" name="__module.model.language_model.layers.2.self_attn/aten::slice/Slice_1" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="731,x1.9">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="526" name="__module.model.language_model.layers.2.self_attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-1" />
<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="739">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="527" name="__module.model.language_model.layers.2.self_attn/aten::mul/Multiply_2" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="740">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="528" name="__module.model.language_model.layers.2.self_attn/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="741,query.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="529" name="Constant_44177" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="530" name="Constant_44170" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="531" name="Constant_44172" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="532" name="Constant_44174" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="533" name="Concat_44175" 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="534" name="Broadcast_44178" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="535" name="ReadValue_43835" type="ReadValue" version="opset6">
<data variable_id="past_key_values.2.keypresent.2.key" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.2.key">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="536" name="Constant_43568" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="537" name="Gather_43569" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="538" name="self.model.language_model.layers.2.self_attn.k_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="270932" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.self_attn.k_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="539" name="__module.model.language_model.layers.2.self_attn.k_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="540" name="__module.model.language_model.layers.2.self_attn.k_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="698">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="541" name="Constant_42108" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="542" name="__module.model.language_model.layers.2.self_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="700">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="543" name="__module.model.language_model.layers.2.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="544" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::to/Convert_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="716,hidden_states.55">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="545" name="Constant_40979" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="546" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::pow/Power_2" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="717">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="547" name="Constant_21555" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="548" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::mean/ReduceMean_1" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="719">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="549" name="Constant_40980" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="550" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="720,mean_squared.13">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="551" name="Constant_40981" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="552" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::pow/Power_3" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="721">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="553" name="__module.model.language_model.layers.2.self_attn.qk_norm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="722,723,k,normed_output.13">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="554" name="__module.model.language_model.layers.2.self_attn/aten::mul/Multiply_3" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="742">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="555" name="Constant_21838" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="556" name="Constant_21840" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8632" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="557" name="Constant_21842" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="558" name="Constant_21794" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="559" name="__module.model.language_model.layers.2.self_attn/aten::slice/Slice_2" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="752,x2">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="560" name="__module.model.language_model.layers.2.self_attn/aten::neg/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="" offset="8648" size="4" />
<output>
<port id="0" precision="I32" />
</output>
</layer>
<layer id="561" name="__module.model.language_model.layers.2.self_attn/aten::neg/ConvertLike_1" type="Convert" version="opset1">
<data destination_type="f32" />
<input>
<port id="0" precision="I32" />
</input>
<output>
<port id="1" precision="FP32" />
</output>
</layer>
<layer id="562" name="__module.model.language_model.layers.2.self_attn/aten::neg/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32" />
</input>
<output>
<port id="2" precision="FP32" names="753">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="563" name="Constant_21735" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="564" name="Constant_21779" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8624" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="565" name="Constant_21781" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8600" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="566" name="Constant_21733" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8640" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="567" name="__module.model.language_model.layers.2.self_attn/aten::slice/Slice_3" type="Slice" version="opset8">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="747,x1">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</output>
</layer>
<layer id="568" name="__module.model.language_model.layers.2.self_attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-1" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>8</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="755">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="569" name="__module.model.language_model.layers.2.self_attn/aten::mul/Multiply_4" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="756">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="570" name="__module.model.language_model.layers.2.self_attn/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="757,key_states.5">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="571" name="__module.model.language_model.layers.2.self_attn/aten::cat/Concat_2" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="767,768,hidden_states.57">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="572" name="__module.model.language_model.layers.2.self_attn/aten::unsqueeze/Unsqueeze_2" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="769,770,771">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="573" name="__module.model.language_model.layers.2.self_attn/aten::expand/Broadcast" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="773,hidden_states.59">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="574" name="Constant_42109" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="575" name="__module.model.language_model.layers.2.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="777,key.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="576" name="Constant_44165" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="577" name="Constant_44158" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="578" name="Constant_44160" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="579" name="Constant_44162" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="580" name="Concat_44163" 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="581" name="Broadcast_44166" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="582" name="ReadValue_43837" type="ReadValue" version="opset6">
<data variable_id="past_key_values.2.valuepresent.2.value" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.2.value">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="583" name="Constant_43571" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="584" name="Gather_43572" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="585" name="self.model.language_model.layers.2.self_attn.v_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="275028" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.self_attn.v_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="586" name="__module.model.language_model.layers.2.self_attn.v_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="587" name="__module.model.language_model.layers.2.self_attn.v_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="703">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="588" name="Constant_42110" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="589" name="__module.model.language_model.layers.2.self_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="705">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="590" name="__module.model.language_model.layers.2.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="591" name="__module.model.language_model.layers.2.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="706,value_states.5">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="592" name="__module.model.language_model.layers.2.self_attn/aten::cat/Concat_3" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="783,784,hidden_states.61">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="593" name="__module.model.language_model.layers.2.self_attn/aten::unsqueeze/Unsqueeze_3" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="785,786,787">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="594" name="__module.model.language_model.layers.2.self_attn/aten::expand/Broadcast_1" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="789,hidden_states.63">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="595" name="Constant_42111" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="596" name="__module.model.language_model.layers.2.self_attn/aten::reshape/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="793,value.5">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="597" name="__module.model.language_model.layers.2.self_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>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="4" precision="FP32" />
</input>
<output>
<port id="5" precision="FP32" names="794,attn_output.13">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="598" name="__module.model.language_model.layers.2.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="599" name="__module.model.language_model.layers.2.self_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>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="795">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="600" name="Constant_42112" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="16952" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="601" name="__module.model.language_model.layers.2.self_attn/aten::reshape/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="798">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="602" name="self.model.language_model.layers.2.self_attn.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="279124" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.self_attn.gate_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="603" name="__module.model.language_model.layers.2.self_attn.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="604" name="__module.model.language_model.layers.2.self_attn.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="801">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="605" name="__module.model.language_model.layers.2.self_attn/aten::sigmoid/Sigmoid" type="Sigmoid" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="802">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="606" name="__module.model.language_model.layers.2.self_attn/aten::mul/Multiply_7" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="803">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="607" name="self.model.language_model.layers.2.self_attn.o_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="287316" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.self_attn.o_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="608" name="__module.model.language_model.layers.2.self_attn.o_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="609" name="__module.model.language_model.layers.2.self_attn.o_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="805,811,x.35,x.37">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="610" name="Constant_40984" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="611" name="Constant_40982" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="612" name="__module.model.language_model.layers.2.post_attention_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="812">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="613" name="Constant_22387" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="614" name="__module.model.language_model.layers.2.post_attention_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="814">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="615" name="Constant_40983" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="616" name="__module.model.language_model.layers.2.post_attention_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="815">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="617" name="__module.model.language_model.layers.2.post_attention_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="618" name="__module.model.language_model.layers.2.post_attention_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="816">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="619" name="__module.model.language_model.layers.2.post_attention_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="817,output.37">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="620" name="Constant_40985" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="621" name="__module.model.language_model.layers.2.post_attention_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="820,821,hidden_states.65,output.39">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="622" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="822,824,x.39,x.41">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="623" name="Constant_40988" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="624" name="Constant_40986" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="625" name="__module.model.language_model.layers.2.pre_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="825">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="626" name="Constant_22452" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="627" name="__module.model.language_model.layers.2.pre_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="827">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="628" name="Constant_40987" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="629" name="__module.model.language_model.layers.2.pre_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="828">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="630" name="__module.model.language_model.layers.2.pre_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="631" name="__module.model.language_model.layers.2.pre_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="829">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="632" name="__module.model.language_model.layers.2.pre_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="830,output.41">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="633" name="Constant_40989" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="634" name="__module.model.language_model.layers.2.pre_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="833,834,output.43">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="635" name="self.model.language_model.layers.2.mlp.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="295508" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.mlp.gate_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="636" name="__module.model.language_model.layers.2.mlp.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="637" name="__module.model.language_model.layers.2.mlp.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="842,input.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="638" name="__module.model.language_model.layers.2.mlp.act_fn/aten::silu/Swish" type="Swish" version="opset4">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="843">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="639" name="self.model.language_model.layers.2.mlp.up_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="328276" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.mlp.up_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="640" name="__module.model.language_model.layers.2.mlp.up_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="641" name="__module.model.language_model.layers.2.mlp.up_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="845">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="642" name="__module.model.language_model.layers.2.mlp/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="846">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="643" name="self.model.language_model.layers.2.mlp.down_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 256" offset="361044" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.2.mlp.down_proj._openvino_orig_weight">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="644" name="__module.model.language_model.layers.2.mlp.down_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="645" name="__module.model.language_model.layers.2.mlp.down_proj/ov_ext::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>256</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="848,850,x.43,x.45">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="646" name="Constant_40992" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="647" name="Constant_40990" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="648" name="__module.model.language_model.layers.2.post_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="851">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="649" name="Constant_22532" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="650" name="__module.model.language_model.layers.2.post_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="853">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="651" name="Constant_40991" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="652" name="__module.model.language_model.layers.2.post_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="854">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="653" name="__module.model.language_model.layers.2.post_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="654" name="__module.model.language_model.layers.2.post_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="855">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="655" name="__module.model.language_model.layers.2.post_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="856,output.45">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="656" name="Constant_40993" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="657" name="__module.model.language_model.layers.2.post_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="859,860,hidden_states.67,output.47">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="658" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="861,873,x.47,x.49">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="659" name="Constant_40996" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="660" name="Constant_40994" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="661" name="__module.model.language_model.layers.3.input_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="874">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="662" name="Constant_22604" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="663" name="__module.model.language_model.layers.3.input_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="876">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="664" name="Constant_40995" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="665" name="__module.model.language_model.layers.3.input_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="877">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="666" name="__module.model.language_model.layers.3.input_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="667" name="__module.model.language_model.layers.3.input_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="878">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="668" name="__module.model.language_model.layers.3.input_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="879,output.49">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="669" name="Constant_40997" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="670" name="__module.model.language_model.layers.3.input_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="882,883,hidden_states.69,output.51">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="671" name="self.model.language_model.layers.3.self_attn.v_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="393812" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.self_attn.v_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="672" name="__module.model.language_model.layers.3.self_attn.v_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="673" name="__module.model.language_model.layers.3.self_attn.v_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="905">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="674" name="Constant_42113" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="675" name="__module.model.language_model.layers.3.self_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="907">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="676" name="__module.model.language_model.layers.3.self_attn/aten::transpose/Constant" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="677" name="__module.model.language_model.layers.3.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="908,value_states">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="678" name="__module.model.language_model.layers.3.self_attn/aten::cat/Concat" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="951,952,hidden_states.83,present.3.value">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="679" name="Constant_44189" type="Const" version="opset1">
<data element_type="f32" shape="" offset="0" size="4" />
<output>
<port id="0" precision="FP32" />
</output>
</layer>
<layer id="680" name="Constant_44182" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="16" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="681" name="Constant_44184" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="8" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="682" name="Constant_44186" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="24" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="683" name="Concat_44187" 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="684" name="Broadcast_44190" 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>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="685" name="ReadValue_43839" type="ReadValue" version="opset6">
<data variable_id="past_key_values.3.keypresent.3.key" variable_type="f32" variable_shape="?,2,?,16" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>0</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="past_key_values.3.key">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="686" name="Constant_43574" type="Const" version="opset1">
<data element_type="i64" shape="" offset="8" size="8" />
<output>
<port id="0" precision="I64" />
</output>
</layer>
<layer id="687" name="Gather_43575" type="Gather" version="opset8">
<data batch_dims="0" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="688" name="self.model.language_model.layers.3.self_attn.k_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="32, 64" offset="397908" size="4096" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.self_attn.k_proj._openvino_orig_weight">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="689" name="__module.model.language_model.layers.3.self_attn.k_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="690" name="__module.model.language_model.layers.3.self_attn.k_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="900">
<dim>-1</dim>
<dim>-1</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="691" name="Constant_42114" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12752" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="692" name="__module.model.language_model.layers.3.self_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="902">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="693" name="__module.model.language_model.layers.3.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="694" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::to/Convert" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>2</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="918,hidden_states.77">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="695" name="Constant_40998" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="696" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="919">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="697" name="Constant_22857" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="698" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="921">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="699" name="Constant_40999" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="700" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="922,mean_squared.17">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="701" name="Constant_41000" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="702" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::pow/Power_1" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</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="FP32" names="923">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="703" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="924,925,key_states,normed_output.17">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="704" name="__module.model.language_model.layers.3.self_attn/aten::cat/Concat_1" type="Concat" version="opset1">
<data axis="-2" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="935,936,hidden_states.79">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="705" name="self.model.language_model.layers.3.self_attn.q_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="402004" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.self_attn.q_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="706" name="__module.model.language_model.layers.3.self_attn.q_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="707" name="__module.model.language_model.layers.3.self_attn.q_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="895">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="708" name="Constant_42115" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="8512" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="709" name="__module.model.language_model.layers.3.self_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>64</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="897">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="710" name="__module.model.language_model.layers.3.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="711" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::to/Convert_1" type="Transpose" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="909,hidden_states.73">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="712" name="Constant_41001" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="713" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::pow/Power_2" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="910">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="714" name="Constant_22811" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="715" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::mean/ReduceMean_1" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="912">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="716" name="Constant_41002" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="717" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::add/Add_1" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="913,mean_squared.15">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="718" name="Constant_41003" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="719" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::pow/Power_3" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</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="FP32" names="914">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="720" name="__module.model.language_model.layers.3.self_attn.qk_norm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="915,916,normed_output.15">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="721" name="Constant_41004" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1, 1" offset="8560" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="722" name="__module.model.language_model.layers.3.self_attn/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</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="FP32" names="917,query">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="723" name="__module.model.language_model.layers.3.self_attn/aten::unsqueeze/Unsqueeze" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="937,938,939">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="724" name="__module.model.language_model.layers.3.self_attn/aten::expand/Broadcast" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="941,hidden_states.81">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="725" name="Constant_42116" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="726" name="__module.model.language_model.layers.3.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="945,key">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="727" name="__module.model.language_model.layers.3.self_attn/aten::unsqueeze/Unsqueeze_1" type="Unsqueeze" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64" />
</input>
<output>
<port id="2" precision="FP32" names="953,954,955">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="728" name="__module.model.language_model.layers.3.self_attn/aten::expand/Broadcast_1" type="Broadcast" version="opset3">
<data mode="bidirectional" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>5</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="957,hidden_states.85">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="729" name="Constant_42117" type="Const" version="opset1">
<data element_type="i64" shape="4" offset="12800" size="32" />
<output>
<port id="0" precision="I64">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="730" name="__module.model.language_model.layers.3.self_attn/aten::reshape/Reshape_1" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="961,value">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="731" name="__module.model.language_model/aten::new_ones/Broadcast_2" type="Const" version="opset1">
<data element_type="boolean" shape="" offset="16928" size="1" />
<output>
<port id="0" precision="BOOL" names="158,result.1" />
</output>
</layer>
<layer id="732" name="__module.model.language_model/aten::__and__/BitwiseAnd_4" 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="161,result.3">
<dim>1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="733" name="__module.model.language_model/aten::__and__/BitwiseAnd_5" 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="165,attention_mask.5">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="734" name="__module.model.language_model/aten::expand/Broadcast_1" 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="167,mask.1">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="735" name="__module.model.language_model/aten::where/Select_1" 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="172,attention_mask_1">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="736" name="__module.model.language_model.layers.3.self_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>16</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="2" precision="FP32">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
<port id="3" precision="FP32">
<dim>-1</dim>
<dim>1</dim>
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="4" precision="FP32" />
</input>
<output>
<port id="5" precision="FP32" names="962,attn_output.19">
<dim>-1</dim>
<dim>4</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="737" name="__module.model.language_model.layers.3.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
<data element_type="i32" shape="4" offset="8544" size="16" />
<output>
<port id="0" precision="I32">
<dim>4</dim>
</port>
</output>
</layer>
<layer id="738" name="__module.model.language_model.layers.3.self_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>16</dim>
</port>
<port id="1" precision="I32">
<dim>4</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="963">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="739" name="Constant_42118" type="Const" version="opset1">
<data element_type="i64" shape="3" offset="16952" size="24" />
<output>
<port id="0" precision="I64">
<dim>3</dim>
</port>
</output>
</layer>
<layer id="740" name="__module.model.language_model.layers.3.self_attn/aten::reshape/Reshape_2" type="Reshape" version="opset1">
<data special_zero="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1" precision="I64">
<dim>3</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="966">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="741" name="self.model.language_model.layers.3.self_attn.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="410196" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.self_attn.gate_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="742" name="__module.model.language_model.layers.3.self_attn.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="743" name="__module.model.language_model.layers.3.self_attn.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="969">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="744" name="__module.model.language_model.layers.3.self_attn/aten::sigmoid/Sigmoid" type="Sigmoid" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="970">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="745" name="__module.model.language_model.layers.3.self_attn/aten::mul/Multiply_3" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="971">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="746" name="self.model.language_model.layers.3.self_attn.o_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 64" offset="418388" size="8192" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.self_attn.o_proj._openvino_orig_weight">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="747" name="__module.model.language_model.layers.3.self_attn.o_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="748" name="__module.model.language_model.layers.3.self_attn.o_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="973,979,x.51,x.53">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="749" name="Constant_41007" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="750" name="Constant_41005" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="751" name="__module.model.language_model.layers.3.post_attention_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="980">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="752" name="Constant_23403" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="753" name="__module.model.language_model.layers.3.post_attention_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="982">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="754" name="Constant_41006" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="755" name="__module.model.language_model.layers.3.post_attention_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="983">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="756" name="__module.model.language_model.layers.3.post_attention_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="757" name="__module.model.language_model.layers.3.post_attention_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="984">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="758" name="__module.model.language_model.layers.3.post_attention_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="985,output.53">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="759" name="Constant_41008" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="760" name="__module.model.language_model.layers.3.post_attention_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="988,989,hidden_states.87,output.55">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="761" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="990,992,x.55,x.57">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="762" name="Constant_41011" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="763" name="Constant_41009" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="764" name="__module.model.language_model.layers.3.pre_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="993">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="765" name="Constant_23468" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="766" name="__module.model.language_model.layers.3.pre_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="995">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="767" name="Constant_41010" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="768" name="__module.model.language_model.layers.3.pre_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="996">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="769" name="__module.model.language_model.layers.3.pre_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="770" name="__module.model.language_model.layers.3.pre_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="997">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="771" name="__module.model.language_model.layers.3.pre_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="998,output.57">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="772" name="Constant_41012" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="773" name="__module.model.language_model.layers.3.pre_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1001,1002,output.59">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="774" name="self.model.language_model.layers.3.mlp.gate_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="426580" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.mlp.gate_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="775" name="__module.model.language_model.layers.3.mlp.gate_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="776" name="__module.model.language_model.layers.3.mlp.gate_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1010,input">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="777" name="__module.model.language_model.layers.3.mlp.act_fn/aten::silu/Swish" type="Swish" version="opset4">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="1011">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="778" name="self.model.language_model.layers.3.mlp.up_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="256, 64" offset="459348" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.mlp.up_proj._openvino_orig_weight">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="779" name="__module.model.language_model.layers.3.mlp.up_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="780" name="__module.model.language_model.layers.3.mlp.up_proj/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>256</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1013">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="781" name="__module.model.language_model.layers.3.mlp/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1014">
<dim>-1</dim>
<dim>-1</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="782" name="self.model.language_model.layers.3.mlp.down_proj._openvino_orig_weight" type="Const" version="opset1">
<data element_type="bf16" shape="64, 256" offset="492116" size="32768" />
<output>
<port id="0" precision="BF16" names="self.model.language_model.layers.3.mlp.down_proj._openvino_orig_weight">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="783" name="__module.model.language_model.layers.3.mlp.down_proj/ov_ext::linear/Convert" type="Convert" version="opset1">
<data destination_type="f32" />
<rt_info>
<attribute name="decompression" version="0" />
</rt_info>
<input>
<port id="0" precision="BF16">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="784" name="__module.model.language_model.layers.3.mlp.down_proj/ov_ext::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>256</dim>
</port>
<port id="1" precision="FP32">
<dim>64</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1016,1018,x,x.59">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="785" name="Constant_41015" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="56" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="786" name="Constant_41013" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="787" name="__module.model.language_model.layers.3.post_feedforward_layernorm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1019">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="788" name="Constant_23548" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="789" name="__module.model.language_model.layers.3.post_feedforward_layernorm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1021">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="790" name="Constant_41014" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="33360" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="791" name="__module.model.language_model.layers.3.post_feedforward_layernorm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1022">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="792" name="__module.model.language_model.layers.3.post_feedforward_layernorm/aten::rsqrt/Sqrt" type="Sqrt" version="opset1">
<input>
<port id="0" precision="FP32">
<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>
</port>
</output>
</layer>
<layer id="793" name="__module.model.language_model.layers.3.post_feedforward_layernorm/aten::rsqrt/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1023">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="794" name="__module.model.language_model.layers.3.post_feedforward_layernorm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1024,output.61">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="795" name="Constant_41016" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 64" offset="60" size="256" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="796" name="__module.model.language_model.layers.3.post_feedforward_layernorm/aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1027,1028,hidden_states.89,output">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="797" name="__module.model.language_model.layers.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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1029,1035,hidden_states.91,hidden_states.93">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="798" name="Constant_41017" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="32" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="799" name="__module.model.language_model.norm/aten::pow/Power" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1036">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="800" name="Constant_23614" type="Const" version="opset1">
<data element_type="i64" shape="1" offset="40" size="8" />
<output>
<port id="0" precision="I64">
<dim>1</dim>
</port>
</output>
</layer>
<layer id="801" name="__module.model.language_model.norm/aten::mean/ReduceMean" type="ReduceMean" version="opset1">
<data keep_dims="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="I64">
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1038">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="802" name="Constant_41018" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="48" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="803" name="__module.model.language_model.norm/aten::add/Add" type="Add" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1039,mean_squared">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="804" name="Constant_41019" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="52" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="805" name="__module.model.language_model.norm/aten::pow/Power_1" type="Power" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1040">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="806" name="__module.model.language_model.norm/aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1041,1043,1044,42,47,52,hidden_states,normed_output,normed_output.19">
<dim>-1</dim>
<dim>-1</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="807" name="self.model.language_model.embed_tokens.weight" type="Const" version="opset1">
<data element_type="f32" shape="202048, 64" offset="524884" size="51724288" />
<output>
<port id="0" precision="FP32" names="self.model.language_model.embed_tokens.weight">
<dim>202048</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="808" name="__module.lm_head/ov_ext::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>64</dim>
</port>
<port id="1" precision="FP32">
<dim>202048</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="1056,logits.1">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
</output>
</layer>
<layer id="809" name="Constant_41020" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="52249172" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="810" name="aten::mul/Multiply" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="55,logits.3">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
</output>
</layer>
<layer id="811" name="Constant_41021" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="52249176" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="812" name="aten::div/Divide" type="Divide" version="opset1">
<data auto_broadcast="numpy" m_pythondiv="true" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="57,logits.5">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
</output>
</layer>
<layer id="813" name="aten::tanh/Tanh" type="Tanh" version="opset1">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
</input>
<output>
<port id="1" precision="FP32" names="58,logits_1">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
</output>
</layer>
<layer id="814" name="Constant_41022" type="Const" version="opset1">
<data element_type="f32" shape="1, 1, 1" offset="52249176" size="4" />
<output>
<port id="0" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="815" name="aten::mul/Multiply_1" type="Multiply" version="opset1">
<data auto_broadcast="numpy" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
<port id="1" precision="FP32">
<dim>1</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="2" precision="FP32" names="logits">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
</output>
</layer>
<layer id="816" name="Assign_43828" type="Assign" version="opset6">
<data variable_id="past_key_values.0.keypresent.0.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="817" name="Assign_43830" type="Assign" version="opset6">
<data variable_id="past_key_values.0.valuepresent.0.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="818" name="Assign_43832" type="Assign" version="opset6">
<data variable_id="past_key_values.1.keypresent.1.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="819" name="Assign_43834" type="Assign" version="opset6">
<data variable_id="past_key_values.1.valuepresent.1.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="820" name="Assign_43836" type="Assign" version="opset6">
<data variable_id="past_key_values.2.keypresent.2.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="821" name="Assign_43838" type="Assign" version="opset6">
<data variable_id="past_key_values.2.valuepresent.2.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="822" name="Assign_43840" type="Assign" version="opset6">
<data variable_id="past_key_values.3.keypresent.3.key" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="823" name="Assign_43842" type="Assign" version="opset6">
<data variable_id="past_key_values.3.valuepresent.3.value" />
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1" precision="FP32">
<dim>-1</dim>
<dim>2</dim>
<dim>-1</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="824" name="Result_23981" type="Result" version="opset1" output_names="logits">
<input>
<port id="0" precision="FP32">
<dim>-1</dim>
<dim>-1</dim>
<dim>202048</dim>
</port>
</input>
</layer>
</layers>
<edges>
<edge from-layer="0" from-port="0" to-layer="204" to-port="0" />
<edge from-layer="1" from-port="0" to-layer="58" to-port="0" />
<edge from-layer="2" from-port="0" to-layer="5" to-port="0" />
<edge from-layer="2" from-port="0" to-layer="25" to-port="0" />
<edge from-layer="2" from-port="0" to-layer="18" to-port="0" />
<edge from-layer="3" from-port="0" to-layer="93" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="161" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="537" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="584" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="687" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="407" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="360" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="16" to-port="1" />
<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="89" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="157" 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="580" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="533" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="356" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="403" to-port="0" />
<edge from-layer="8" from-port="3" to-layer="683" 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="678" to-port="0" />
<edge from-layer="17" from-port="0" to-layer="18" to-port="1" />
<edge from-layer="18" from-port="2" to-layer="20" to-port="0" />
<edge from-layer="19" from-port="0" to-layer="20" to-port="1" />
<edge from-layer="20" from-port="2" to-layer="22" to-port="0" />
<edge from-layer="21" from-port="0" to-layer="22" to-port="1" />
<edge from-layer="22" from-port="2" to-layer="24" to-port="0" />
<edge from-layer="23" from-port="0" to-layer="24" to-port="1" />
<edge from-layer="24" from-port="2" to-layer="25" to-port="1" />
<edge from-layer="25" from-port="2" to-layer="268" to-port="0" />
<edge from-layer="25" from-port="2" to-layer="28" to-port="0" />
<edge from-layer="25" from-port="2" to-layer="35" to-port="0" />
<edge from-layer="26" from-port="0" to-layer="34" to-port="0" />
<edge from-layer="27" from-port="0" to-layer="28" to-port="1" />
<edge from-layer="28" from-port="2" to-layer="30" to-port="0" />
<edge from-layer="29" from-port="0" to-layer="30" to-port="1" />
<edge from-layer="30" from-port="2" to-layer="32" to-port="0" />
<edge from-layer="31" from-port="0" to-layer="32" to-port="1" />
<edge from-layer="32" from-port="2" to-layer="33" to-port="0" />
<edge from-layer="33" from-port="1" to-layer="34" to-port="1" />
<edge from-layer="34" from-port="2" to-layer="35" to-port="1" />
<edge from-layer="35" from-port="2" to-layer="37" to-port="0" />
<edge from-layer="36" from-port="0" to-layer="37" to-port="1" />
<edge from-layer="37" from-port="2" to-layer="164" to-port="0" />
<edge from-layer="37" from-port="2" to-layer="96" to-port="0" />
<edge from-layer="37" from-port="2" to-layer="141" to-port="0" />
<edge from-layer="37" from-port="2" to-layer="40" to-port="0" />
<edge from-layer="37" from-port="2" to-layer="250" to-port="0" />
<edge from-layer="38" from-port="0" to-layer="39" to-port="0" />
<edge from-layer="39" from-port="1" to-layer="40" to-port="1" />
<edge from-layer="40" from-port="2" to-layer="42" to-port="0" />
<edge from-layer="41" from-port="0" to-layer="42" to-port="1" />
<edge from-layer="42" from-port="2" to-layer="44" to-port="0" />
<edge from-layer="43" from-port="0" to-layer="44" to-port="1" />
<edge from-layer="44" from-port="2" to-layer="53" to-port="0" />
<edge from-layer="44" from-port="2" to-layer="46" to-port="0" />
<edge from-layer="45" from-port="0" to-layer="46" to-port="1" />
<edge from-layer="46" from-port="2" to-layer="48" to-port="0" />
<edge from-layer="47" from-port="0" to-layer="48" to-port="1" />
<edge from-layer="48" from-port="2" to-layer="50" to-port="0" />
<edge from-layer="49" from-port="0" to-layer="50" to-port="1" />
<edge from-layer="50" from-port="2" to-layer="52" to-port="0" />
<edge from-layer="51" from-port="0" to-layer="52" to-port="1" />
<edge from-layer="52" from-port="2" to-layer="53" to-port="1" />
<edge from-layer="53" from-port="2" to-layer="55" to-port="0" />
<edge from-layer="54" from-port="0" to-layer="55" to-port="1" />
<edge from-layer="55" from-port="2" to-layer="71" to-port="0" />
<edge from-layer="55" from-port="2" to-layer="79" to-port="0" />
<edge from-layer="55" from-port="2" to-layer="66" to-port="0" />
<edge from-layer="56" from-port="0" to-layer="60" to-port="0" />
<edge from-layer="57" from-port="0" to-layer="182" to-port="1" />
<edge from-layer="57" from-port="0" to-layer="218" to-port="1" />
<edge from-layer="57" from-port="0" to-layer="58" to-port="1" />
<edge from-layer="57" from-port="0" to-layer="194" to-port="1" />
<edge from-layer="57" from-port="0" to-layer="65" to-port="1" />
<edge from-layer="57" from-port="0" to-layer="82" 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="60" to-port="1" />
<edge from-layer="60" from-port="2" to-layer="62" to-port="0" />
<edge from-layer="61" from-port="0" to-layer="62" to-port="1" />
<edge from-layer="62" from-port="2" to-layer="63" to-port="0" />
<edge from-layer="62" from-port="2" to-layer="63" to-port="1" />
<edge from-layer="63" from-port="2" to-layer="81" to-port="0" />
<edge from-layer="63" from-port="2" to-layer="64" to-port="0" />
<edge from-layer="64" from-port="1" to-layer="65" to-port="0" />
<edge from-layer="65" from-port="2" to-layer="554" to-port="1" />
<edge from-layer="65" from-port="2" to-layer="66" to-port="1" />
<edge from-layer="65" from-port="2" to-layer="335" to-port="1" />
<edge from-layer="65" from-port="2" to-layer="110" to-port="1" />
<edge from-layer="65" from-port="2" to-layer="512" to-port="1" />
<edge from-layer="65" from-port="2" to-layer="377" to-port="1" />
<edge from-layer="66" from-port="2" to-layer="84" to-port="0" />
<edge from-layer="67" from-port="0" to-layer="71" to-port="1" />
<edge from-layer="68" from-port="0" to-layer="71" to-port="2" />
<edge from-layer="69" from-port="0" to-layer="71" to-port="3" />
<edge from-layer="70" from-port="0" to-layer="71" to-port="4" />
<edge from-layer="71" from-port="5" to-layer="74" to-port="0" />
<edge from-layer="72" from-port="0" to-layer="73" to-port="0" />
<edge from-layer="73" from-port="1" to-layer="74" to-port="1" />
<edge from-layer="74" from-port="2" to-layer="80" to-port="0" />
<edge from-layer="75" from-port="0" to-layer="79" to-port="1" />
<edge from-layer="76" from-port="0" to-layer="79" to-port="2" />
<edge from-layer="77" from-port="0" to-layer="79" to-port="3" />
<edge from-layer="78" from-port="0" to-layer="79" to-port="4" />
<edge from-layer="79" from-port="5" to-layer="80" to-port="1" />
<edge from-layer="80" from-port="2" to-layer="83" to-port="0" />
<edge from-layer="81" from-port="1" to-layer="82" to-port="0" />
<edge from-layer="82" from-port="2" to-layer="569" to-port="1" />
<edge from-layer="82" from-port="2" to-layer="350" to-port="1" />
<edge from-layer="82" from-port="2" to-layer="392" to-port="1" />
<edge from-layer="82" from-port="2" to-layer="527" to-port="1" />
<edge from-layer="82" from-port="2" to-layer="83" to-port="1" />
<edge from-layer="82" from-port="2" to-layer="125" to-port="1" />
<edge from-layer="83" from-port="2" to-layer="84" to-port="1" />
<edge from-layer="84" from-port="2" to-layer="243" to-port="0" />
<edge from-layer="85" from-port="0" to-layer="90" to-port="0" />
<edge from-layer="86" from-port="0" to-layer="89" to-port="1" />
<edge from-layer="87" from-port="0" to-layer="89" to-port="2" />
<edge from-layer="88" from-port="0" to-layer="89" to-port="3" />
<edge from-layer="89" from-port="4" to-layer="90" to-port="1" />
<edge from-layer="90" from-port="2" to-layer="91" to-port="0" />
<edge from-layer="91" from-port="1" to-layer="93" to-port="0" />
<edge from-layer="92" from-port="0" to-layer="93" to-port="2" />
<edge from-layer="93" from-port="3" to-layer="127" to-port="0" />
<edge from-layer="93" from-port="3" to-layer="135" to-port="0" />
<edge from-layer="94" from-port="0" to-layer="95" to-port="0" />
<edge from-layer="95" from-port="1" to-layer="96" to-port="1" />
<edge from-layer="96" from-port="2" to-layer="98" to-port="0" />
<edge from-layer="97" from-port="0" to-layer="98" to-port="1" />
<edge from-layer="98" from-port="2" to-layer="100" 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="102" to-port="0" />
<edge from-layer="100" from-port="2" to-layer="109" to-port="0" />
<edge from-layer="101" from-port="0" to-layer="102" to-port="1" />
<edge from-layer="102" from-port="2" 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="108" to-port="0" />
<edge from-layer="107" from-port="0" to-layer="108" to-port="1" />
<edge from-layer="108" from-port="2" to-layer="109" to-port="1" />
<edge from-layer="109" from-port="2" to-layer="110" to-port="0" />
<edge from-layer="109" from-port="2" to-layer="115" to-port="0" />
<edge from-layer="109" from-port="2" to-layer="123" to-port="0" />
<edge from-layer="110" from-port="2" to-layer="126" to-port="0" />
<edge from-layer="111" from-port="0" to-layer="115" to-port="1" />
<edge from-layer="112" from-port="0" to-layer="115" to-port="2" />
<edge from-layer="113" from-port="0" to-layer="115" to-port="3" />
<edge from-layer="114" from-port="0" to-layer="115" to-port="4" />
<edge from-layer="115" from-port="5" to-layer="118" to-port="0" />
<edge from-layer="116" from-port="0" to-layer="117" to-port="0" />
<edge from-layer="117" from-port="1" to-layer="118" to-port="1" />
<edge from-layer="118" from-port="2" to-layer="124" to-port="0" />
<edge from-layer="119" from-port="0" to-layer="123" to-port="1" />
<edge from-layer="120" from-port="0" to-layer="123" to-port="2" />
<edge from-layer="121" from-port="0" to-layer="123" to-port="3" />
<edge from-layer="122" from-port="0" to-layer="123" to-port="4" />
<edge from-layer="123" from-port="5" to-layer="124" to-port="1" />
<edge from-layer="124" from-port="2" to-layer="125" to-port="0" />
<edge from-layer="125" from-port="2" to-layer="126" to-port="1" />
<edge from-layer="126" from-port="2" to-layer="127" to-port="1" />
<edge from-layer="127" from-port="2" to-layer="129" to-port="0" />
<edge from-layer="127" from-port="2" to-layer="816" to-port="0" />
<edge from-layer="127" from-port="2" to-layer="130" to-port="0" />
<edge from-layer="128" from-port="0" to-layer="416" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="395" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="572" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="170" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="727" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="723" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="183" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="593" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="219" to-port="1" />
<edge from-layer="128" from-port="0" to-layer="129" to-port="1" />
<edge from-layer="129" from-port="2" to-layer="150" to-port="0" />
<edge from-layer="130" from-port="1" to-layer="133" to-port="0" />
<edge from-layer="131" from-port="0" to-layer="133" to-port="1" />
<edge from-layer="132" from-port="0" to-layer="133" to-port="2" />
<edge from-layer="133" from-port="3" to-layer="149" to-port="0" />
<edge from-layer="134" from-port="0" to-layer="149" to-port="1" />
<edge from-layer="135" from-port="1" to-layer="138" to-port="0" />
<edge from-layer="136" from-port="0" to-layer="138" to-port="1" />
<edge from-layer="137" from-port="0" to-layer="138" to-port="2" />
<edge from-layer="138" from-port="3" to-layer="177" to-port="0" />
<edge from-layer="138" from-port="3" to-layer="190" to-port="0" />
<edge from-layer="138" from-port="3" to-layer="140" to-port="0" />
<edge from-layer="139" from-port="0" to-layer="140" to-port="1" />
<edge from-layer="140" from-port="2" to-layer="147" to-port="0" />
<edge from-layer="141" from-port="1" to-layer="215" to-port="0" />
<edge from-layer="141" from-port="1" to-layer="144" to-port="0" />
<edge from-layer="142" from-port="0" to-layer="144" to-port="1" />
<edge from-layer="143" from-port="0" to-layer="144" to-port="2" />
<edge from-layer="144" from-port="3" to-layer="146" to-port="0" />
<edge from-layer="144" from-port="3" to-layer="177" to-port="1" />
<edge from-layer="145" from-port="0" to-layer="146" to-port="1" />
<edge from-layer="146" from-port="2" to-layer="147" to-port="1" />
<edge from-layer="146" from-port="2" to-layer="186" to-port="0" />
<edge from-layer="146" from-port="2" to-layer="237" to-port="2" />
<edge from-layer="147" from-port="2" to-layer="149" to-port="2" />
<edge from-layer="148" from-port="0" to-layer="149" to-port="3" />
<edge from-layer="149" from-port="4" to-layer="396" to-port="1" />
<edge from-layer="149" from-port="4" to-layer="573" to-port="1" />
<edge from-layer="149" from-port="4" to-layer="594" to-port="1" />
<edge from-layer="149" from-port="4" to-layer="150" to-port="1" />
<edge from-layer="149" from-port="4" to-layer="724" to-port="1" />
<edge from-layer="149" from-port="4" to-layer="728" to-port="1" />
<edge from-layer="149" from-port="4" to-layer="171" to-port="1" />
<edge from-layer="149" from-port="4" to-layer="417" to-port="1" />
<edge from-layer="150" from-port="2" to-layer="152" to-port="0" />
<edge from-layer="151" from-port="0" to-layer="152" to-port="1" />
<edge from-layer="152" from-port="2" to-layer="243" to-port="1" />
<edge from-layer="153" from-port="0" to-layer="158" to-port="0" />
<edge from-layer="154" from-port="0" to-layer="157" to-port="1" />
<edge from-layer="155" from-port="0" to-layer="157" to-port="2" />
<edge from-layer="156" from-port="0" to-layer="157" to-port="3" />
<edge from-layer="157" from-port="4" to-layer="158" to-port="1" />
<edge from-layer="158" from-port="2" to-layer="159" to-port="0" />
<edge from-layer="159" from-port="1" to-layer="161" to-port="0" />
<edge from-layer="160" from-port="0" to-layer="161" to-port="2" />
<edge from-layer="161" from-port="3" to-layer="169" to-port="0" />
<edge from-layer="162" from-port="0" to-layer="163" to-port="0" />
<edge from-layer="163" from-port="1" to-layer="164" to-port="1" />
<edge from-layer="164" from-port="2" to-layer="166" to-port="0" />
<edge from-layer="165" from-port="0" to-layer="166" to-port="1" />
<edge from-layer="166" from-port="2" to-layer="168" to-port="0" />
<edge from-layer="167" from-port="0" to-layer="168" to-port="1" />
<edge from-layer="168" from-port="2" to-layer="169" to-port="1" />
<edge from-layer="169" from-port="2" to-layer="817" to-port="0" />
<edge from-layer="169" from-port="2" to-layer="170" to-port="0" />
<edge from-layer="170" from-port="2" to-layer="171" to-port="0" />
<edge from-layer="171" from-port="2" to-layer="173" to-port="0" />
<edge from-layer="172" from-port="0" to-layer="173" to-port="1" />
<edge from-layer="173" from-port="2" to-layer="243" to-port="2" />
<edge from-layer="174" from-port="0" to-layer="203" to-port="0" />
<edge from-layer="175" from-port="0" to-layer="200" to-port="0" />
<edge from-layer="176" from-port="0" to-layer="179" to-port="0" />
<edge from-layer="177" from-port="2" to-layer="190" to-port="1" />
<edge from-layer="177" from-port="2" to-layer="236" to-port="0" />
<edge from-layer="177" from-port="2" to-layer="179" to-port="1" />
<edge from-layer="178" from-port="0" to-layer="179" to-port="2" />
<edge from-layer="179" from-port="3" to-layer="181" to-port="0" />
<edge from-layer="180" from-port="0" to-layer="193" to-port="1" />
<edge from-layer="180" from-port="0" to-layer="191" to-port="1" />
<edge from-layer="180" from-port="0" to-layer="191" to-port="2" />
<edge from-layer="180" from-port="0" to-layer="181" to-port="1" />
<edge from-layer="181" from-port="2" to-layer="182" to-port="0" />
<edge from-layer="182" from-port="2" to-layer="183" to-port="0" />
<edge from-layer="183" from-port="2" to-layer="199" to-port="0" />
<edge from-layer="183" from-port="2" to-layer="201" to-port="0" />
<edge from-layer="183" from-port="2" to-layer="211" to-port="0" />
<edge from-layer="184" from-port="0" to-layer="188" to-port="0" />
<edge from-layer="185" from-port="0" to-layer="186" to-port="1" />
<edge from-layer="186" from-port="2" to-layer="188" to-port="1" />
<edge from-layer="187" from-port="0" to-layer="188" to-port="2" />
<edge from-layer="188" from-port="3" to-layer="192" to-port="0" />
<edge from-layer="189" from-port="0" to-layer="190" to-port="2" />
<edge from-layer="190" from-port="3" to-layer="191" to-port="0" />
<edge from-layer="191" from-port="3" to-layer="192" to-port="1" />
<edge from-layer="192" from-port="2" to-layer="193" to-port="0" />
<edge from-layer="193" from-port="2" to-layer="194" to-port="0" />
<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="195" from-port="0" to-layer="220" to-port="1" />
<edge from-layer="196" from-port="2" to-layer="198" to-port="0" />
<edge from-layer="196" from-port="2" to-layer="201" to-port="1" />
<edge from-layer="197" from-port="0" to-layer="198" to-port="1" />
<edge from-layer="198" from-port="2" to-layer="199" to-port="1" />
<edge from-layer="199" from-port="2" 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="2" to-layer="732" to-port="1" />
<edge from-layer="201" from-port="2" to-layer="202" to-port="1" />
<edge from-layer="202" from-port="2" to-layer="203" to-port="1" />
<edge from-layer="203" from-port="2" to-layer="231" to-port="0" />
<edge from-layer="204" from-port="1" to-layer="210" to-port="0" />
<edge from-layer="204" from-port="1" to-layer="205" to-port="0" />
<edge from-layer="205" from-port="1" to-layer="223" to-port="0" />
<edge from-layer="205" from-port="1" 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="1" />
<edge from-layer="210" from-port="2" to-layer="226" to-port="0" />
<edge from-layer="211" from-port="1" to-layer="225" to-port="0" />
<edge from-layer="212" from-port="0" to-layer="217" to-port="0" />
<edge from-layer="213" from-port="0" to-layer="215" to-port="1" />
<edge from-layer="214" from-port="0" to-layer="215" to-port="2" />
<edge from-layer="215" from-port="3" to-layer="233" to-port="0" />
<edge from-layer="215" from-port="3" to-layer="217" to-port="1" />
<edge from-layer="216" from-port="0" to-layer="217" to-port="2" />
<edge from-layer="217" from-port="3" to-layer="218" to-port="0" />
<edge from-layer="218" from-port="2" to-layer="219" to-port="0" />
<edge from-layer="219" from-port="2" to-layer="220" to-port="0" />
<edge from-layer="220" from-port="2" to-layer="221" to-port="0" />
<edge from-layer="221" from-port="1" to-layer="224" to-port="0" />
<edge from-layer="222" from-port="0" to-layer="223" to-port="1" />
<edge from-layer="222" from-port="0" to-layer="226" to-port="2" />
<edge from-layer="223" from-port="3" to-layer="224" to-port="1" />
<edge from-layer="224" from-port="2" to-layer="225" to-port="1" />
<edge from-layer="225" from-port="2" to-layer="226" to-port="1" />
<edge from-layer="225" from-port="2" to-layer="229" to-port="0" />
<edge from-layer="226" from-port="3" to-layer="228" to-port="0" />
<edge from-layer="227" from-port="0" to-layer="228" to-port="1" />
<edge from-layer="228" from-port="2" to-layer="230" to-port="0" />
<edge from-layer="229" from-port="1" to-layer="230" to-port="1" />
<edge from-layer="230" from-port="2" to-layer="733" to-port="1" />
<edge from-layer="230" from-port="2" to-layer="231" to-port="1" />
<edge from-layer="231" from-port="2" to-layer="238" 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="237" to-port="0" />
<edge from-layer="234" from-port="0" to-layer="237" to-port="1" />
<edge from-layer="235" from-port="0" to-layer="236" to-port="1" />
<edge from-layer="236" from-port="2" to-layer="237" to-port="3" />
<edge from-layer="237" from-port="4" to-layer="238" to-port="1" />
<edge from-layer="237" from-port="4" to-layer="734" to-port="1" />
<edge from-layer="238" from-port="2" to-layer="241" to-port="0" />
<edge from-layer="239" from-port="0" to-layer="241" to-port="1" />
<edge from-layer="239" from-port="0" to-layer="735" to-port="1" />
<edge from-layer="240" from-port="0" to-layer="735" to-port="2" />
<edge from-layer="240" from-port="0" to-layer="241" to-port="2" />
<edge from-layer="241" from-port="3" to-layer="597" to-port="3" />
<edge from-layer="241" from-port="3" to-layer="420" to-port="3" />
<edge from-layer="241" from-port="3" to-layer="243" to-port="3" />
<edge from-layer="242" from-port="0" to-layer="420" to-port="4" />
<edge from-layer="242" from-port="0" to-layer="597" to-port="4" />
<edge from-layer="242" from-port="0" to-layer="736" to-port="4" />
<edge from-layer="242" from-port="0" to-layer="243" to-port="4" />
<edge from-layer="243" from-port="5" to-layer="245" to-port="0" />
<edge from-layer="244" from-port="0" to-layer="245" to-port="1" />
<edge from-layer="245" from-port="2" to-layer="247" 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="252" to-port="0" />
<edge from-layer="248" from-port="0" to-layer="249" to-port="0" />
<edge from-layer="249" from-port="1" to-layer="250" to-port="1" />
<edge from-layer="250" from-port="2" to-layer="251" to-port="0" />
<edge from-layer="251" from-port="1" to-layer="252" to-port="1" />
<edge from-layer="252" from-port="2" to-layer="255" to-port="0" />
<edge from-layer="253" from-port="0" to-layer="254" to-port="0" />
<edge from-layer="254" from-port="1" to-layer="255" to-port="1" />
<edge from-layer="255" from-port="2" to-layer="265" to-port="0" />
<edge from-layer="255" from-port="2" to-layer="258" to-port="0" />
<edge from-layer="256" from-port="0" to-layer="264" to-port="0" />
<edge from-layer="257" from-port="0" to-layer="258" to-port="1" />
<edge from-layer="258" from-port="2" to-layer="260" 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="264" to-port="1" />
<edge from-layer="264" from-port="2" to-layer="265" to-port="1" />
<edge from-layer="265" from-port="2" to-layer="267" to-port="0" />
<edge from-layer="266" from-port="0" to-layer="267" to-port="1" />
<edge from-layer="267" from-port="2" to-layer="268" to-port="1" />
<edge from-layer="268" from-port="2" to-layer="271" to-port="0" />
<edge from-layer="268" from-port="2" to-layer="278" to-port="0" />
<edge from-layer="268" from-port="2" to-layer="304" to-port="0" />
<edge from-layer="269" from-port="0" to-layer="277" to-port="0" />
<edge from-layer="270" from-port="0" to-layer="271" to-port="1" />
<edge from-layer="271" from-port="2" 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="276" to-port="0" />
<edge from-layer="276" from-port="1" to-layer="277" to-port="1" />
<edge from-layer="277" from-port="2" to-layer="278" to-port="1" />
<edge from-layer="278" from-port="2" to-layer="280" to-port="0" />
<edge from-layer="279" from-port="0" to-layer="280" to-port="1" />
<edge from-layer="280" from-port="2" to-layer="287" to-port="0" />
<edge from-layer="280" from-port="2" to-layer="283" to-port="0" />
<edge from-layer="281" from-port="0" to-layer="282" to-port="0" />
<edge from-layer="282" from-port="1" to-layer="283" to-port="1" />
<edge from-layer="283" from-port="2" to-layer="284" to-port="0" />
<edge from-layer="284" from-port="1" to-layer="288" to-port="0" />
<edge from-layer="285" from-port="0" to-layer="286" to-port="0" />
<edge from-layer="286" from-port="1" to-layer="287" to-port="1" />
<edge from-layer="287" from-port="2" to-layer="288" to-port="1" />
<edge from-layer="288" from-port="2" to-layer="291" to-port="0" />
<edge from-layer="289" from-port="0" to-layer="290" to-port="0" />
<edge from-layer="290" from-port="1" to-layer="291" to-port="1" />
<edge from-layer="291" from-port="2" to-layer="294" to-port="0" />
<edge from-layer="291" from-port="2" to-layer="301" to-port="0" />
<edge from-layer="292" from-port="0" to-layer="300" to-port="0" />
<edge from-layer="293" from-port="0" to-layer="294" to-port="1" />
<edge from-layer="294" from-port="2" to-layer="296" to-port="0" />
<edge from-layer="295" from-port="0" to-layer="296" to-port="1" />
<edge from-layer="296" from-port="2" to-layer="298" to-port="0" />
<edge from-layer="297" from-port="0" to-layer="298" to-port="1" />
<edge from-layer="298" from-port="2" to-layer="299" to-port="0" />
<edge from-layer="299" from-port="1" to-layer="300" to-port="1" />
<edge from-layer="300" from-port="2" to-layer="301" to-port="1" />
<edge from-layer="301" from-port="2" to-layer="303" to-port="0" />
<edge from-layer="302" from-port="0" to-layer="303" to-port="1" />
<edge from-layer="303" from-port="2" to-layer="304" to-port="1" />
<edge from-layer="304" from-port="2" to-layer="445" to-port="0" />
<edge from-layer="304" from-port="2" to-layer="314" to-port="0" />
<edge from-layer="304" from-port="2" to-layer="307" to-port="0" />
<edge from-layer="305" from-port="0" to-layer="313" to-port="0" />
<edge from-layer="306" from-port="0" to-layer="307" to-port="1" />
<edge from-layer="307" from-port="2" to-layer="309" to-port="0" />
<edge from-layer="308" from-port="0" to-layer="309" to-port="1" />
<edge from-layer="309" from-port="2" to-layer="311" to-port="0" />
<edge from-layer="310" from-port="0" to-layer="311" to-port="1" />
<edge from-layer="311" from-port="2" to-layer="312" to-port="0" />
<edge from-layer="312" from-port="1" to-layer="313" to-port="1" />
<edge from-layer="313" from-port="2" to-layer="314" to-port="1" />
<edge from-layer="314" from-port="2" to-layer="316" to-port="0" />
<edge from-layer="315" from-port="0" to-layer="316" to-port="1" />
<edge from-layer="316" from-port="2" to-layer="363" to-port="0" />
<edge from-layer="316" from-port="2" to-layer="427" to-port="0" />
<edge from-layer="316" from-port="2" to-layer="410" to-port="0" />
<edge from-layer="316" from-port="2" to-layer="319" to-port="0" />
<edge from-layer="317" from-port="0" to-layer="318" to-port="0" />
<edge from-layer="318" from-port="1" to-layer="319" 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="332" to-port="0" />
<edge from-layer="323" from-port="2" to-layer="325" to-port="0" />
<edge from-layer="324" from-port="0" to-layer="325" to-port="1" />
<edge from-layer="325" from-port="2" to-layer="327" to-port="0" />
<edge from-layer="326" from-port="0" to-layer="327" to-port="1" />
<edge from-layer="327" from-port="2" to-layer="329" to-port="0" />
<edge from-layer="328" from-port="0" to-layer="329" to-port="1" />
<edge from-layer="329" from-port="2" to-layer="331" to-port="0" />
<edge from-layer="330" from-port="0" to-layer="331" to-port="1" />
<edge from-layer="331" from-port="2" to-layer="332" to-port="1" />
<edge from-layer="332" from-port="2" to-layer="334" to-port="0" />
<edge from-layer="333" from-port="0" to-layer="334" to-port="1" />
<edge from-layer="334" from-port="2" to-layer="335" to-port="0" />
<edge from-layer="334" from-port="2" to-layer="340" to-port="0" />
<edge from-layer="334" from-port="2" to-layer="348" to-port="0" />
<edge from-layer="335" from-port="2" to-layer="351" to-port="0" />
<edge from-layer="336" from-port="0" to-layer="340" to-port="1" />
<edge from-layer="337" from-port="0" to-layer="340" to-port="2" />
<edge from-layer="338" from-port="0" to-layer="340" to-port="3" />
<edge from-layer="339" from-port="0" to-layer="340" to-port="4" />
<edge from-layer="340" from-port="5" to-layer="343" to-port="0" />
<edge from-layer="341" from-port="0" to-layer="342" to-port="0" />
<edge from-layer="342" from-port="1" to-layer="343" to-port="1" />
<edge from-layer="343" from-port="2" to-layer="349" to-port="0" />
<edge from-layer="344" from-port="0" to-layer="348" to-port="1" />
<edge from-layer="345" from-port="0" to-layer="348" to-port="2" />
<edge from-layer="346" from-port="0" to-layer="348" to-port="3" />
<edge from-layer="347" from-port="0" to-layer="348" to-port="4" />
<edge from-layer="348" from-port="5" to-layer="349" to-port="1" />
<edge from-layer="349" from-port="2" to-layer="350" to-port="0" />
<edge from-layer="350" from-port="2" to-layer="351" to-port="1" />
<edge from-layer="351" from-port="2" to-layer="420" to-port="0" />
<edge from-layer="352" from-port="0" to-layer="357" to-port="0" />
<edge from-layer="353" from-port="0" to-layer="356" to-port="1" />
<edge from-layer="354" from-port="0" to-layer="356" to-port="2" />
<edge from-layer="355" from-port="0" to-layer="356" to-port="3" />
<edge from-layer="356" from-port="4" 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="1" to-layer="360" to-port="0" />
<edge from-layer="359" from-port="0" to-layer="360" to-port="2" />
<edge from-layer="360" from-port="3" to-layer="394" to-port="0" />
<edge from-layer="361" from-port="0" to-layer="362" to-port="0" />
<edge from-layer="362" from-port="1" to-layer="363" to-port="1" />
<edge from-layer="363" from-port="2" to-layer="365" to-port="0" />
<edge from-layer="364" from-port="0" to-layer="365" to-port="1" />
<edge from-layer="365" from-port="2" to-layer="367" 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="376" to-port="0" />
<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="371" to-port="0" />
<edge from-layer="370" from-port="0" to-layer="371" to-port="1" />
<edge from-layer="371" from-port="2" to-layer="373" to-port="0" />
<edge from-layer="372" from-port="0" 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="376" to-port="1" />
<edge from-layer="376" from-port="2" to-layer="377" to-port="0" />
<edge from-layer="376" from-port="2" to-layer="390" to-port="0" />
<edge from-layer="376" from-port="2" to-layer="382" to-port="0" />
<edge from-layer="377" from-port="2" to-layer="393" to-port="0" />
<edge from-layer="378" from-port="0" to-layer="382" to-port="1" />
<edge from-layer="379" from-port="0" to-layer="382" to-port="2" />
<edge from-layer="380" from-port="0" to-layer="382" to-port="3" />
<edge from-layer="381" from-port="0" to-layer="382" to-port="4" />
<edge from-layer="382" from-port="5" to-layer="385" to-port="0" />
<edge from-layer="383" from-port="0" to-layer="384" to-port="0" />
<edge from-layer="384" from-port="1" to-layer="385" to-port="1" />
<edge from-layer="385" from-port="2" to-layer="391" to-port="0" />
<edge from-layer="386" from-port="0" to-layer="390" to-port="1" />
<edge from-layer="387" from-port="0" to-layer="390" to-port="2" />
<edge from-layer="388" from-port="0" to-layer="390" to-port="3" />
<edge from-layer="389" from-port="0" to-layer="390" to-port="4" />
<edge from-layer="390" from-port="5" to-layer="391" to-port="1" />
<edge from-layer="391" from-port="2" to-layer="392" to-port="0" />
<edge from-layer="392" from-port="2" to-layer="393" to-port="1" />
<edge from-layer="393" from-port="2" to-layer="394" to-port="1" />
<edge from-layer="394" from-port="2" to-layer="818" to-port="0" />
<edge from-layer="394" from-port="2" to-layer="395" to-port="0" />
<edge from-layer="395" from-port="2" to-layer="396" to-port="0" />
<edge from-layer="396" from-port="2" to-layer="398" to-port="0" />
<edge from-layer="397" from-port="0" to-layer="398" to-port="1" />
<edge from-layer="398" from-port="2" to-layer="420" to-port="1" />
<edge from-layer="399" from-port="0" to-layer="404" to-port="0" />
<edge from-layer="400" from-port="0" to-layer="403" to-port="1" />
<edge from-layer="401" from-port="0" to-layer="403" to-port="2" />
<edge from-layer="402" from-port="0" to-layer="403" to-port="3" />
<edge from-layer="403" from-port="4" to-layer="404" to-port="1" />
<edge from-layer="404" from-port="2" to-layer="405" to-port="0" />
<edge from-layer="405" from-port="1" to-layer="407" to-port="0" />
<edge from-layer="406" from-port="0" to-layer="407" to-port="2" />
<edge from-layer="407" from-port="3" to-layer="415" to-port="0" />
<edge from-layer="408" from-port="0" to-layer="409" to-port="0" />
<edge from-layer="409" from-port="1" to-layer="410" to-port="1" />
<edge from-layer="410" from-port="2" to-layer="412" to-port="0" />
<edge from-layer="411" from-port="0" to-layer="412" to-port="1" />
<edge from-layer="412" from-port="2" to-layer="414" to-port="0" />
<edge from-layer="413" from-port="0" to-layer="414" to-port="1" />
<edge from-layer="414" from-port="2" to-layer="415" to-port="1" />
<edge from-layer="415" from-port="2" to-layer="416" to-port="0" />
<edge from-layer="415" from-port="2" to-layer="819" to-port="0" />
<edge from-layer="416" from-port="2" to-layer="417" to-port="0" />
<edge from-layer="417" from-port="2" to-layer="419" to-port="0" />
<edge from-layer="418" from-port="0" to-layer="419" to-port="1" />
<edge from-layer="419" from-port="2" to-layer="420" to-port="2" />
<edge from-layer="420" from-port="5" to-layer="422" to-port="0" />
<edge from-layer="421" from-port="0" to-layer="422" to-port="1" />
<edge from-layer="422" from-port="2" to-layer="424" to-port="0" />
<edge from-layer="423" from-port="0" to-layer="424" to-port="1" />
<edge from-layer="424" from-port="2" to-layer="429" to-port="0" />
<edge from-layer="425" from-port="0" to-layer="426" to-port="0" />
<edge from-layer="426" from-port="1" to-layer="427" to-port="1" />
<edge from-layer="427" from-port="2" to-layer="428" to-port="0" />
<edge from-layer="428" from-port="1" to-layer="429" to-port="1" />
<edge from-layer="429" from-port="2" to-layer="432" to-port="0" />
<edge from-layer="430" from-port="0" to-layer="431" to-port="0" />
<edge from-layer="431" from-port="1" to-layer="432" to-port="1" />
<edge from-layer="432" from-port="2" to-layer="435" to-port="0" />
<edge from-layer="432" from-port="2" to-layer="442" to-port="0" />
<edge from-layer="433" from-port="0" to-layer="441" to-port="0" />
<edge from-layer="434" from-port="0" to-layer="435" to-port="1" />
<edge from-layer="435" from-port="2" to-layer="437" to-port="0" />
<edge from-layer="436" from-port="0" to-layer="437" to-port="1" />
<edge from-layer="437" from-port="2" to-layer="439" to-port="0" />
<edge from-layer="438" from-port="0" to-layer="439" to-port="1" />
<edge from-layer="439" from-port="2" to-layer="440" to-port="0" />
<edge from-layer="440" from-port="1" to-layer="441" to-port="1" />
<edge from-layer="441" from-port="2" to-layer="442" to-port="1" />
<edge from-layer="442" from-port="2" to-layer="444" to-port="0" />
<edge from-layer="443" from-port="0" to-layer="444" to-port="1" />
<edge from-layer="444" from-port="2" to-layer="445" to-port="1" />
<edge from-layer="445" from-port="2" to-layer="448" to-port="0" />
<edge from-layer="445" from-port="2" to-layer="455" to-port="0" />
<edge from-layer="445" from-port="2" to-layer="481" to-port="0" />
<edge from-layer="446" from-port="0" to-layer="454" to-port="0" />
<edge from-layer="447" from-port="0" to-layer="448" to-port="1" />
<edge from-layer="448" from-port="2" to-layer="450" to-port="0" />
<edge from-layer="449" from-port="0" to-layer="450" to-port="1" />
<edge from-layer="450" from-port="2" to-layer="452" to-port="0" />
<edge from-layer="451" from-port="0" to-layer="452" to-port="1" />
<edge from-layer="452" from-port="2" to-layer="453" to-port="0" />
<edge from-layer="453" from-port="1" to-layer="454" to-port="1" />
<edge from-layer="454" from-port="2" to-layer="455" to-port="1" />
<edge from-layer="455" from-port="2" to-layer="457" to-port="0" />
<edge from-layer="456" from-port="0" to-layer="457" to-port="1" />
<edge from-layer="457" from-port="2" to-layer="460" to-port="0" />
<edge from-layer="457" from-port="2" to-layer="464" to-port="0" />
<edge from-layer="458" from-port="0" to-layer="459" to-port="0" />
<edge from-layer="459" from-port="1" to-layer="460" to-port="1" />
<edge from-layer="460" from-port="2" to-layer="461" to-port="0" />
<edge from-layer="461" from-port="1" to-layer="465" to-port="0" />
<edge from-layer="462" from-port="0" to-layer="463" to-port="0" />
<edge from-layer="463" from-port="1" to-layer="464" to-port="1" />
<edge from-layer="464" from-port="2" to-layer="465" to-port="1" />
<edge from-layer="465" from-port="2" to-layer="468" to-port="0" />
<edge from-layer="466" from-port="0" to-layer="467" to-port="0" />
<edge from-layer="467" from-port="1" to-layer="468" to-port="1" />
<edge from-layer="468" from-port="2" to-layer="471" to-port="0" />
<edge from-layer="468" from-port="2" to-layer="478" to-port="0" />
<edge from-layer="469" from-port="0" to-layer="477" to-port="0" />
<edge from-layer="470" from-port="0" to-layer="471" to-port="1" />
<edge from-layer="471" from-port="2" to-layer="473" to-port="0" />
<edge from-layer="472" from-port="0" to-layer="473" to-port="1" />
<edge from-layer="473" from-port="2" to-layer="475" to-port="0" />
<edge from-layer="474" from-port="0" to-layer="475" to-port="1" />
<edge from-layer="475" from-port="2" to-layer="476" to-port="0" />
<edge from-layer="476" from-port="1" to-layer="477" to-port="1" />
<edge from-layer="477" from-port="2" to-layer="478" to-port="1" />
<edge from-layer="478" from-port="2" to-layer="480" to-port="0" />
<edge from-layer="479" from-port="0" to-layer="480" to-port="1" />
<edge from-layer="480" from-port="2" to-layer="481" to-port="1" />
<edge from-layer="481" from-port="2" to-layer="484" to-port="0" />
<edge from-layer="481" from-port="2" to-layer="491" to-port="0" />
<edge from-layer="481" from-port="2" to-layer="622" to-port="0" />
<edge from-layer="482" from-port="0" to-layer="490" to-port="0" />
<edge from-layer="483" from-port="0" to-layer="484" to-port="1" />
<edge from-layer="484" from-port="2" to-layer="486" to-port="0" />
<edge from-layer="485" from-port="0" to-layer="486" to-port="1" />
<edge from-layer="486" from-port="2" to-layer="488" to-port="0" />
<edge from-layer="487" from-port="0" to-layer="488" to-port="1" />
<edge from-layer="488" from-port="2" to-layer="489" to-port="0" />
<edge from-layer="489" from-port="1" to-layer="490" to-port="1" />
<edge from-layer="490" from-port="2" to-layer="491" to-port="1" />
<edge from-layer="491" from-port="2" to-layer="493" to-port="0" />
<edge from-layer="492" from-port="0" to-layer="493" to-port="1" />
<edge from-layer="493" from-port="2" to-layer="496" to-port="0" />
<edge from-layer="493" from-port="2" to-layer="540" to-port="0" />
<edge from-layer="493" from-port="2" to-layer="604" to-port="0" />
<edge from-layer="493" from-port="2" to-layer="587" to-port="0" />
<edge from-layer="494" from-port="0" to-layer="495" to-port="0" />
<edge from-layer="495" from-port="1" to-layer="496" to-port="1" />
<edge from-layer="496" from-port="2" to-layer="498" to-port="0" />
<edge from-layer="497" from-port="0" to-layer="498" to-port="1" />
<edge from-layer="498" from-port="2" to-layer="500" to-port="0" />
<edge from-layer="499" from-port="0" to-layer="500" to-port="1" />
<edge from-layer="500" from-port="2" to-layer="509" to-port="0" />
<edge from-layer="500" from-port="2" to-layer="502" to-port="0" />
<edge from-layer="501" from-port="0" to-layer="502" to-port="1" />
<edge from-layer="502" from-port="2" to-layer="504" to-port="0" />
<edge from-layer="503" from-port="0" to-layer="504" to-port="1" />
<edge from-layer="504" from-port="2" to-layer="506" to-port="0" />
<edge from-layer="505" from-port="0" to-layer="506" to-port="1" />
<edge from-layer="506" from-port="2" to-layer="508" to-port="0" />
<edge from-layer="507" from-port="0" to-layer="508" to-port="1" />
<edge from-layer="508" from-port="2" to-layer="509" to-port="1" />
<edge from-layer="509" from-port="2" to-layer="511" to-port="0" />
<edge from-layer="510" from-port="0" to-layer="511" to-port="1" />
<edge from-layer="511" from-port="2" to-layer="512" to-port="0" />
<edge from-layer="511" from-port="2" to-layer="517" to-port="0" />
<edge from-layer="511" from-port="2" to-layer="525" to-port="0" />
<edge from-layer="512" from-port="2" to-layer="528" to-port="0" />
<edge from-layer="513" from-port="0" to-layer="517" to-port="1" />
<edge from-layer="514" from-port="0" to-layer="517" to-port="2" />
<edge from-layer="515" from-port="0" to-layer="517" to-port="3" />
<edge from-layer="516" from-port="0" to-layer="517" to-port="4" />
<edge from-layer="517" from-port="5" to-layer="520" to-port="0" />
<edge from-layer="518" from-port="0" to-layer="519" to-port="0" />
<edge from-layer="519" from-port="1" to-layer="520" to-port="1" />
<edge from-layer="520" from-port="2" to-layer="526" to-port="0" />
<edge from-layer="521" from-port="0" to-layer="525" to-port="1" />
<edge from-layer="522" from-port="0" to-layer="525" to-port="2" />
<edge from-layer="523" from-port="0" to-layer="525" to-port="3" />
<edge from-layer="524" from-port="0" to-layer="525" to-port="4" />
<edge from-layer="525" from-port="5" to-layer="526" to-port="1" />
<edge from-layer="526" from-port="2" to-layer="527" to-port="0" />
<edge from-layer="527" from-port="2" to-layer="528" to-port="1" />
<edge from-layer="528" from-port="2" to-layer="597" to-port="0" />
<edge from-layer="529" from-port="0" to-layer="534" to-port="0" />
<edge from-layer="530" from-port="0" to-layer="533" to-port="1" />
<edge from-layer="531" from-port="0" to-layer="533" to-port="2" />
<edge from-layer="532" from-port="0" to-layer="533" to-port="3" />
<edge from-layer="533" from-port="4" to-layer="534" to-port="1" />
<edge from-layer="534" from-port="2" to-layer="535" to-port="0" />
<edge from-layer="535" from-port="1" to-layer="537" to-port="0" />
<edge from-layer="536" from-port="0" to-layer="537" to-port="2" />
<edge from-layer="537" from-port="3" to-layer="571" to-port="0" />
<edge from-layer="538" from-port="0" to-layer="539" to-port="0" />
<edge from-layer="539" from-port="1" to-layer="540" to-port="1" />
<edge from-layer="540" from-port="2" to-layer="542" to-port="0" />
<edge from-layer="541" from-port="0" to-layer="542" to-port="1" />
<edge from-layer="542" from-port="2" to-layer="544" to-port="0" />
<edge from-layer="543" from-port="0" to-layer="544" to-port="1" />
<edge from-layer="544" from-port="2" to-layer="546" to-port="0" />
<edge from-layer="544" from-port="2" to-layer="553" to-port="0" />
<edge from-layer="545" from-port="0" to-layer="546" to-port="1" />
<edge from-layer="546" from-port="2" to-layer="548" to-port="0" />
<edge from-layer="547" from-port="0" to-layer="548" to-port="1" />
<edge from-layer="548" from-port="2" to-layer="550" to-port="0" />
<edge from-layer="549" from-port="0" to-layer="550" to-port="1" />
<edge from-layer="550" from-port="2" to-layer="552" to-port="0" />
<edge from-layer="551" from-port="0" to-layer="552" to-port="1" />
<edge from-layer="552" from-port="2" to-layer="553" to-port="1" />
<edge from-layer="553" from-port="2" to-layer="554" to-port="0" />
<edge from-layer="553" from-port="2" to-layer="559" to-port="0" />
<edge from-layer="553" from-port="2" to-layer="567" to-port="0" />
<edge from-layer="554" from-port="2" to-layer="570" to-port="0" />
<edge from-layer="555" from-port="0" to-layer="559" to-port="1" />
<edge from-layer="556" from-port="0" to-layer="559" to-port="2" />
<edge from-layer="557" from-port="0" to-layer="559" to-port="3" />
<edge from-layer="558" from-port="0" to-layer="559" to-port="4" />
<edge from-layer="559" from-port="5" to-layer="562" to-port="0" />
<edge from-layer="560" from-port="0" to-layer="561" to-port="0" />
<edge from-layer="561" from-port="1" to-layer="562" to-port="1" />
<edge from-layer="562" from-port="2" to-layer="568" to-port="0" />
<edge from-layer="563" from-port="0" to-layer="567" to-port="1" />
<edge from-layer="564" from-port="0" to-layer="567" to-port="2" />
<edge from-layer="565" from-port="0" to-layer="567" to-port="3" />
<edge from-layer="566" from-port="0" to-layer="567" to-port="4" />
<edge from-layer="567" from-port="5" to-layer="568" to-port="1" />
<edge from-layer="568" from-port="2" to-layer="569" to-port="0" />
<edge from-layer="569" from-port="2" to-layer="570" to-port="1" />
<edge from-layer="570" from-port="2" to-layer="571" to-port="1" />
<edge from-layer="571" from-port="2" to-layer="572" to-port="0" />
<edge from-layer="571" from-port="2" to-layer="820" to-port="0" />
<edge from-layer="572" from-port="2" to-layer="573" to-port="0" />
<edge from-layer="573" from-port="2" to-layer="575" to-port="0" />
<edge from-layer="574" from-port="0" to-layer="575" to-port="1" />
<edge from-layer="575" from-port="2" to-layer="597" to-port="1" />
<edge from-layer="576" from-port="0" to-layer="581" to-port="0" />
<edge from-layer="577" from-port="0" to-layer="580" to-port="1" />
<edge from-layer="578" from-port="0" to-layer="580" to-port="2" />
<edge from-layer="579" from-port="0" to-layer="580" to-port="3" />
<edge from-layer="580" from-port="4" to-layer="581" to-port="1" />
<edge from-layer="581" from-port="2" to-layer="582" to-port="0" />
<edge from-layer="582" from-port="1" to-layer="584" to-port="0" />
<edge from-layer="583" from-port="0" to-layer="584" to-port="2" />
<edge from-layer="584" from-port="3" to-layer="592" to-port="0" />
<edge from-layer="585" from-port="0" to-layer="586" to-port="0" />
<edge from-layer="586" from-port="1" to-layer="587" to-port="1" />
<edge from-layer="587" from-port="2" to-layer="589" to-port="0" />
<edge from-layer="588" from-port="0" to-layer="589" to-port="1" />
<edge from-layer="589" from-port="2" to-layer="591" to-port="0" />
<edge from-layer="590" from-port="0" to-layer="591" to-port="1" />
<edge from-layer="591" from-port="2" to-layer="592" to-port="1" />
<edge from-layer="592" from-port="2" to-layer="821" to-port="0" />
<edge from-layer="592" from-port="2" to-layer="593" to-port="0" />
<edge from-layer="593" from-port="2" to-layer="594" to-port="0" />
<edge from-layer="594" from-port="2" to-layer="596" to-port="0" />
<edge from-layer="595" from-port="0" to-layer="596" to-port="1" />
<edge from-layer="596" from-port="2" to-layer="597" to-port="2" />
<edge from-layer="597" from-port="5" to-layer="599" to-port="0" />
<edge from-layer="598" from-port="0" to-layer="599" to-port="1" />
<edge from-layer="599" from-port="2" to-layer="601" to-port="0" />
<edge from-layer="600" from-port="0" to-layer="601" to-port="1" />
<edge from-layer="601" from-port="2" to-layer="606" to-port="0" />
<edge from-layer="602" from-port="0" to-layer="603" to-port="0" />
<edge from-layer="603" from-port="1" to-layer="604" to-port="1" />
<edge from-layer="604" from-port="2" to-layer="605" to-port="0" />
<edge from-layer="605" from-port="1" to-layer="606" to-port="1" />
<edge from-layer="606" from-port="2" to-layer="609" to-port="0" />
<edge from-layer="607" from-port="0" to-layer="608" to-port="0" />
<edge from-layer="608" from-port="1" to-layer="609" to-port="1" />
<edge from-layer="609" from-port="2" to-layer="612" to-port="0" />
<edge from-layer="609" from-port="2" to-layer="619" to-port="0" />
<edge from-layer="610" from-port="0" to-layer="618" to-port="0" />
<edge from-layer="611" from-port="0" to-layer="612" to-port="1" />
<edge from-layer="612" from-port="2" to-layer="614" to-port="0" />
<edge from-layer="613" from-port="0" to-layer="614" to-port="1" />
<edge from-layer="614" from-port="2" to-layer="616" to-port="0" />
<edge from-layer="615" from-port="0" to-layer="616" to-port="1" />
<edge from-layer="616" from-port="2" to-layer="617" to-port="0" />
<edge from-layer="617" from-port="1" to-layer="618" to-port="1" />
<edge from-layer="618" from-port="2" to-layer="619" to-port="1" />
<edge from-layer="619" from-port="2" to-layer="621" to-port="0" />
<edge from-layer="620" from-port="0" to-layer="621" to-port="1" />
<edge from-layer="621" from-port="2" to-layer="622" to-port="1" />
<edge from-layer="622" from-port="2" to-layer="658" to-port="0" />
<edge from-layer="622" from-port="2" to-layer="632" to-port="0" />
<edge from-layer="622" from-port="2" to-layer="625" to-port="0" />
<edge from-layer="623" from-port="0" to-layer="631" to-port="0" />
<edge from-layer="624" from-port="0" to-layer="625" to-port="1" />
<edge from-layer="625" from-port="2" to-layer="627" to-port="0" />
<edge from-layer="626" from-port="0" to-layer="627" to-port="1" />
<edge from-layer="627" from-port="2" to-layer="629" to-port="0" />
<edge from-layer="628" from-port="0" to-layer="629" to-port="1" />
<edge from-layer="629" from-port="2" to-layer="630" to-port="0" />
<edge from-layer="630" from-port="1" to-layer="631" to-port="1" />
<edge from-layer="631" from-port="2" to-layer="632" to-port="1" />
<edge from-layer="632" from-port="2" to-layer="634" to-port="0" />
<edge from-layer="633" from-port="0" to-layer="634" to-port="1" />
<edge from-layer="634" from-port="2" to-layer="641" to-port="0" />
<edge from-layer="634" from-port="2" to-layer="637" to-port="0" />
<edge from-layer="635" from-port="0" to-layer="636" to-port="0" />
<edge from-layer="636" from-port="1" to-layer="637" to-port="1" />
<edge from-layer="637" from-port="2" to-layer="638" to-port="0" />
<edge from-layer="638" from-port="1" to-layer="642" to-port="0" />
<edge from-layer="639" from-port="0" to-layer="640" to-port="0" />
<edge from-layer="640" from-port="1" to-layer="641" to-port="1" />
<edge from-layer="641" from-port="2" to-layer="642" to-port="1" />
<edge from-layer="642" from-port="2" to-layer="645" to-port="0" />
<edge from-layer="643" from-port="0" to-layer="644" to-port="0" />
<edge from-layer="644" from-port="1" to-layer="645" to-port="1" />
<edge from-layer="645" from-port="2" to-layer="655" to-port="0" />
<edge from-layer="645" from-port="2" to-layer="648" to-port="0" />
<edge from-layer="646" from-port="0" to-layer="654" to-port="0" />
<edge from-layer="647" from-port="0" to-layer="648" to-port="1" />
<edge from-layer="648" from-port="2" to-layer="650" to-port="0" />
<edge from-layer="649" from-port="0" to-layer="650" to-port="1" />
<edge from-layer="650" from-port="2" to-layer="652" to-port="0" />
<edge from-layer="651" from-port="0" to-layer="652" to-port="1" />
<edge from-layer="652" from-port="2" to-layer="653" to-port="0" />
<edge from-layer="653" from-port="1" to-layer="654" to-port="1" />
<edge from-layer="654" from-port="2" to-layer="655" to-port="1" />
<edge from-layer="655" from-port="2" to-layer="657" to-port="0" />
<edge from-layer="656" from-port="0" to-layer="657" to-port="1" />
<edge from-layer="657" from-port="2" to-layer="658" to-port="1" />
<edge from-layer="658" from-port="2" to-layer="668" to-port="0" />
<edge from-layer="658" from-port="2" to-layer="661" to-port="0" />
<edge from-layer="658" from-port="2" to-layer="761" to-port="0" />
<edge from-layer="659" from-port="0" to-layer="667" to-port="0" />
<edge from-layer="660" from-port="0" to-layer="661" to-port="1" />
<edge from-layer="661" from-port="2" to-layer="663" to-port="0" />
<edge from-layer="662" from-port="0" to-layer="663" to-port="1" />
<edge from-layer="663" from-port="2" to-layer="665" to-port="0" />
<edge from-layer="664" from-port="0" to-layer="665" to-port="1" />
<edge from-layer="665" from-port="2" to-layer="666" to-port="0" />
<edge from-layer="666" from-port="1" to-layer="667" to-port="1" />
<edge from-layer="667" from-port="2" to-layer="668" to-port="1" />
<edge from-layer="668" from-port="2" to-layer="670" to-port="0" />
<edge from-layer="669" from-port="0" to-layer="670" to-port="1" />
<edge from-layer="670" from-port="2" to-layer="690" to-port="0" />
<edge from-layer="670" from-port="2" to-layer="707" to-port="0" />
<edge from-layer="670" from-port="2" to-layer="743" to-port="0" />
<edge from-layer="670" from-port="2" to-layer="673" to-port="0" />
<edge from-layer="671" from-port="0" to-layer="672" to-port="0" />
<edge from-layer="672" from-port="1" to-layer="673" to-port="1" />
<edge from-layer="673" from-port="2" to-layer="675" to-port="0" />
<edge from-layer="674" from-port="0" to-layer="675" to-port="1" />
<edge from-layer="675" from-port="2" to-layer="677" to-port="0" />
<edge from-layer="676" from-port="0" to-layer="677" to-port="1" />
<edge from-layer="677" from-port="2" to-layer="678" to-port="1" />
<edge from-layer="678" from-port="2" to-layer="727" to-port="0" />
<edge from-layer="678" from-port="2" to-layer="823" to-port="0" />
<edge from-layer="679" from-port="0" to-layer="684" to-port="0" />
<edge from-layer="680" from-port="0" to-layer="683" to-port="1" />
<edge from-layer="681" from-port="0" to-layer="683" to-port="2" />
<edge from-layer="682" from-port="0" to-layer="683" to-port="3" />
<edge from-layer="683" from-port="4" to-layer="684" to-port="1" />
<edge from-layer="684" from-port="2" to-layer="685" to-port="0" />
<edge from-layer="685" from-port="1" to-layer="687" to-port="0" />
<edge from-layer="686" from-port="0" to-layer="687" to-port="2" />
<edge from-layer="687" from-port="3" to-layer="704" to-port="0" />
<edge from-layer="688" from-port="0" to-layer="689" to-port="0" />
<edge from-layer="689" from-port="1" to-layer="690" to-port="1" />
<edge from-layer="690" from-port="2" to-layer="692" to-port="0" />
<edge from-layer="691" from-port="0" to-layer="692" to-port="1" />
<edge from-layer="692" from-port="2" to-layer="694" to-port="0" />
<edge from-layer="693" from-port="0" to-layer="694" to-port="1" />
<edge from-layer="694" from-port="2" to-layer="703" to-port="0" />
<edge from-layer="694" from-port="2" to-layer="696" to-port="0" />
<edge from-layer="695" from-port="0" to-layer="696" to-port="1" />
<edge from-layer="696" from-port="2" to-layer="698" to-port="0" />
<edge from-layer="697" from-port="0" to-layer="698" to-port="1" />
<edge from-layer="698" from-port="2" to-layer="700" to-port="0" />
<edge from-layer="699" from-port="0" to-layer="700" to-port="1" />
<edge from-layer="700" from-port="2" to-layer="702" to-port="0" />
<edge from-layer="701" from-port="0" to-layer="702" to-port="1" />
<edge from-layer="702" from-port="2" to-layer="703" to-port="1" />
<edge from-layer="703" from-port="2" to-layer="704" to-port="1" />
<edge from-layer="704" from-port="2" to-layer="723" to-port="0" />
<edge from-layer="704" from-port="2" to-layer="822" to-port="0" />
<edge from-layer="705" from-port="0" to-layer="706" to-port="0" />
<edge from-layer="706" from-port="1" to-layer="707" to-port="1" />
<edge from-layer="707" from-port="2" to-layer="709" to-port="0" />
<edge from-layer="708" from-port="0" to-layer="709" to-port="1" />
<edge from-layer="709" from-port="2" to-layer="711" to-port="0" />
<edge from-layer="710" from-port="0" to-layer="711" to-port="1" />
<edge from-layer="711" from-port="2" to-layer="720" to-port="0" />
<edge from-layer="711" from-port="2" to-layer="713" to-port="0" />
<edge from-layer="712" from-port="0" to-layer="713" to-port="1" />
<edge from-layer="713" from-port="2" to-layer="715" to-port="0" />
<edge from-layer="714" from-port="0" to-layer="715" to-port="1" />
<edge from-layer="715" from-port="2" to-layer="717" to-port="0" />
<edge from-layer="716" from-port="0" to-layer="717" to-port="1" />
<edge from-layer="717" from-port="2" to-layer="719" to-port="0" />
<edge from-layer="718" from-port="0" to-layer="719" to-port="1" />
<edge from-layer="719" from-port="2" to-layer="720" to-port="1" />
<edge from-layer="720" from-port="2" to-layer="722" to-port="0" />
<edge from-layer="721" from-port="0" to-layer="722" to-port="1" />
<edge from-layer="722" from-port="2" to-layer="736" to-port="0" />
<edge from-layer="723" from-port="2" to-layer="724" to-port="0" />
<edge from-layer="724" from-port="2" to-layer="726" to-port="0" />
<edge from-layer="725" from-port="0" to-layer="726" to-port="1" />
<edge from-layer="726" from-port="2" to-layer="736" to-port="1" />
<edge from-layer="727" from-port="2" to-layer="728" to-port="0" />
<edge from-layer="728" from-port="2" to-layer="730" to-port="0" />
<edge from-layer="729" from-port="0" to-layer="730" to-port="1" />
<edge from-layer="730" from-port="2" to-layer="736" to-port="2" />
<edge from-layer="731" from-port="0" to-layer="732" to-port="0" />
<edge from-layer="732" from-port="2" to-layer="733" to-port="0" />
<edge from-layer="733" from-port="2" to-layer="734" to-port="0" />
<edge from-layer="734" from-port="2" to-layer="735" to-port="0" />
<edge from-layer="735" from-port="3" to-layer="736" to-port="3" />
<edge from-layer="736" from-port="5" to-layer="738" to-port="0" />
<edge from-layer="737" from-port="0" to-layer="738" to-port="1" />
<edge from-layer="738" from-port="2" to-layer="740" to-port="0" />
<edge from-layer="739" from-port="0" to-layer="740" to-port="1" />
<edge from-layer="740" from-port="2" to-layer="745" to-port="0" />
<edge from-layer="741" from-port="0" to-layer="742" to-port="0" />
<edge from-layer="742" from-port="1" to-layer="743" to-port="1" />
<edge from-layer="743" from-port="2" to-layer="744" to-port="0" />
<edge from-layer="744" from-port="1" to-layer="745" to-port="1" />
<edge from-layer="745" from-port="2" to-layer="748" to-port="0" />
<edge from-layer="746" from-port="0" to-layer="747" to-port="0" />
<edge from-layer="747" from-port="1" to-layer="748" to-port="1" />
<edge from-layer="748" from-port="2" to-layer="758" to-port="0" />
<edge from-layer="748" from-port="2" to-layer="751" to-port="0" />
<edge from-layer="749" from-port="0" to-layer="757" to-port="0" />
<edge from-layer="750" from-port="0" to-layer="751" to-port="1" />
<edge from-layer="751" from-port="2" to-layer="753" to-port="0" />
<edge from-layer="752" from-port="0" to-layer="753" to-port="1" />
<edge from-layer="753" from-port="2" to-layer="755" to-port="0" />
<edge from-layer="754" from-port="0" to-layer="755" to-port="1" />
<edge from-layer="755" from-port="2" to-layer="756" to-port="0" />
<edge from-layer="756" from-port="1" to-layer="757" to-port="1" />
<edge from-layer="757" from-port="2" to-layer="758" to-port="1" />
<edge from-layer="758" from-port="2" to-layer="760" to-port="0" />
<edge from-layer="759" from-port="0" to-layer="760" to-port="1" />
<edge from-layer="760" from-port="2" to-layer="761" to-port="1" />
<edge from-layer="761" from-port="2" to-layer="797" to-port="0" />
<edge from-layer="761" from-port="2" to-layer="771" to-port="0" />
<edge from-layer="761" from-port="2" to-layer="764" to-port="0" />
<edge from-layer="762" from-port="0" to-layer="770" to-port="0" />
<edge from-layer="763" from-port="0" to-layer="764" to-port="1" />
<edge from-layer="764" from-port="2" to-layer="766" to-port="0" />
<edge from-layer="765" from-port="0" to-layer="766" to-port="1" />
<edge from-layer="766" from-port="2" to-layer="768" to-port="0" />
<edge from-layer="767" from-port="0" to-layer="768" to-port="1" />
<edge from-layer="768" from-port="2" to-layer="769" to-port="0" />
<edge from-layer="769" from-port="1" to-layer="770" to-port="1" />
<edge from-layer="770" from-port="2" to-layer="771" to-port="1" />
<edge from-layer="771" from-port="2" to-layer="773" to-port="0" />
<edge from-layer="772" from-port="0" to-layer="773" to-port="1" />
<edge from-layer="773" from-port="2" to-layer="780" to-port="0" />
<edge from-layer="773" from-port="2" to-layer="776" to-port="0" />
<edge from-layer="774" from-port="0" to-layer="775" to-port="0" />
<edge from-layer="775" from-port="1" to-layer="776" to-port="1" />
<edge from-layer="776" from-port="2" to-layer="777" to-port="0" />
<edge from-layer="777" from-port="1" to-layer="781" to-port="0" />
<edge from-layer="778" from-port="0" to-layer="779" to-port="0" />
<edge from-layer="779" from-port="1" to-layer="780" to-port="1" />
<edge from-layer="780" from-port="2" to-layer="781" to-port="1" />
<edge from-layer="781" from-port="2" to-layer="784" to-port="0" />
<edge from-layer="782" from-port="0" to-layer="783" to-port="0" />
<edge from-layer="783" from-port="1" to-layer="784" to-port="1" />
<edge from-layer="784" from-port="2" to-layer="794" to-port="0" />
<edge from-layer="784" from-port="2" to-layer="787" to-port="0" />
<edge from-layer="785" from-port="0" to-layer="793" to-port="0" />
<edge from-layer="786" from-port="0" to-layer="787" to-port="1" />
<edge from-layer="787" from-port="2" to-layer="789" to-port="0" />
<edge from-layer="788" from-port="0" to-layer="789" to-port="1" />
<edge from-layer="789" from-port="2" to-layer="791" to-port="0" />
<edge from-layer="790" from-port="0" to-layer="791" to-port="1" />
<edge from-layer="791" from-port="2" to-layer="792" to-port="0" />
<edge from-layer="792" from-port="1" to-layer="793" to-port="1" />
<edge from-layer="793" from-port="2" to-layer="794" to-port="1" />
<edge from-layer="794" from-port="2" to-layer="796" to-port="0" />
<edge from-layer="795" from-port="0" to-layer="796" to-port="1" />
<edge from-layer="796" from-port="2" to-layer="797" to-port="1" />
<edge from-layer="797" from-port="2" to-layer="806" to-port="0" />
<edge from-layer="797" from-port="2" to-layer="799" to-port="0" />
<edge from-layer="798" from-port="0" to-layer="799" to-port="1" />
<edge from-layer="799" from-port="2" to-layer="801" to-port="0" />
<edge from-layer="800" from-port="0" to-layer="801" to-port="1" />
<edge from-layer="801" from-port="2" to-layer="803" to-port="0" />
<edge from-layer="802" from-port="0" to-layer="803" to-port="1" />
<edge from-layer="803" from-port="2" to-layer="805" to-port="0" />
<edge from-layer="804" from-port="0" to-layer="805" to-port="1" />
<edge from-layer="805" from-port="2" to-layer="806" to-port="1" />
<edge from-layer="806" from-port="2" to-layer="808" to-port="0" />
<edge from-layer="807" from-port="0" to-layer="808" to-port="1" />
<edge from-layer="808" from-port="2" to-layer="810" to-port="0" />
<edge from-layer="809" from-port="0" to-layer="810" to-port="1" />
<edge from-layer="810" from-port="2" to-layer="812" to-port="0" />
<edge from-layer="811" from-port="0" to-layer="812" to-port="1" />
<edge from-layer="812" from-port="2" to-layer="813" to-port="0" />
<edge from-layer="813" from-port="1" to-layer="815" to-port="0" />
<edge from-layer="814" from-port="0" to-layer="815" to-port="1" />
<edge from-layer="815" from-port="2" to-layer="824" to-port="0" />
</edges>
<rt_info>
<info name="OpenVINO Runtime" value="2026.5.0-23697-32776832f8d" />
<Runtime_version value="2026.5.0-23697-32776832f8d" />
<conversion_parameters>
<framework value="pytorch" />
<is_python_object value="True" />
</conversion_parameters>
<hidden_states_decoder_layers value="{&quot;layers&quot;:{&quot;0&quot;:{&quot;producer&quot;:&quot;__module.model.language_model.layers.0/aten::add/Add_1&quot;,&quot;output_index&quot;:0},&quot;1&quot;:{&quot;producer&quot;:&quot;__module.model.language_model.layers.1/aten::add/Add_1&quot;,&quot;output_index&quot;:0},&quot;2&quot;:{&quot;producer&quot;:&quot;__module.model.language_model.layers.2/aten::add/Add_1&quot;,&quot;output_index&quot;:0},&quot;3&quot;:{&quot;producer&quot;:&quot;__module.model.language_model.norm/aten::mul/Multiply&quot;,&quot;output_index&quot;:0}}}" />
<optimum>
<optimum_intel_version value="2.3.0.dev0+80ea8c3b" />
<optimum_version value="2.3.0" />
<pytorch_version value="2.13.0+cpu" />
<transformers_version value="5.15.1" />
</optimum>
<runtime_options>
<ACTIVATIONS_SCALE_FACTOR value="8.0" />
<KV_CACHE_PRECISION value="f16" />
</runtime_options>
</rt_info>
</net>