diff --git "a/openvino_encoder_model.xml" "b/openvino_encoder_model.xml"
--- "a/openvino_encoder_model.xml"
+++ "b/openvino_encoder_model.xml"
@@ -11,37 +11,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="1" name="self.conv1.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 80, 3" offset="0" size="245760" />
+		<layer id="1" name="self.conv1.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 80, 3" offset="0" size="491520" />
 			<output>
-				<port id="0" precision="FP16" names="self.conv1.weight">
+				<port id="0" precision="FP32" names="self.conv1.weight">
 					<dim>512</dim>
 					<dim>80</dim>
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="2" name="self.conv1.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>80</dim>
-					<dim>3</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>80</dim>
-					<dim>3</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="3" name="__module.conv1/aten::_convolution/Convolution" type="Convolution" version="opset1">
+		<layer id="2" name="__module.conv1/aten::_convolution/Convolution" type="Convolution" version="opset1">
 			<data strides="1" dilations="1" pads_begin="1" pads_end="1" auto_pad="explicit" />
 			<input>
 				<port id="0" precision="FP32">
@@ -63,37 +43,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="4" name="__module.conv1/aten::_convolution/Reshape_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 512, 1" offset="245760" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>512</dim>
-					<dim>1</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="5" name="__module.conv1/aten::_convolution/Reshape" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>512</dim>
-					<dim>1</dim>
-				</port>
-			</input>
+		<layer id="3" name="__module.conv1/aten::_convolution/Reshape" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 512, 1" offset="491520" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>512</dim>
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="6" name="__module.conv1/aten::_convolution/Add" type="Add" version="opset1">
+		<layer id="4" name="__module.conv1/aten::_convolution/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -115,7 +75,7 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="7" name="aten::gelu/Gelu" type="Gelu" version="opset7">
+		<layer id="5" name="aten::gelu/Gelu" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -132,37 +92,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="8" name="self.conv2.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512, 3" offset="246784" size="1572864" />
-			<output>
-				<port id="0" precision="FP16" names="self.conv2.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-					<dim>3</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="9" name="self.conv2.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-					<dim>3</dim>
-				</port>
-			</input>
+		<layer id="6" name="self.conv2.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512, 3" offset="493568" size="3145728" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.conv2.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="10" name="__module.conv2/aten::_convolution/Convolution" type="Convolution" version="opset1">
+		<layer id="7" name="__module.conv2/aten::_convolution/Convolution" type="Convolution" version="opset1">
 			<data strides="2" dilations="1" pads_begin="1" pads_end="1" auto_pad="explicit" />
 			<input>
 				<port id="0" precision="FP32">
@@ -184,37 +124,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="11" name="__module.conv2/aten::_convolution/Reshape_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 512, 1" offset="1819648" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>512</dim>
-					<dim>1</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="12" name="__module.conv2/aten::_convolution/Reshape" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>512</dim>
-					<dim>1</dim>
-				</port>
-			</input>
+		<layer id="8" name="__module.conv2/aten::_convolution/Reshape" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 512, 1" offset="3639296" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>512</dim>
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="13" name="__module.conv2/aten::_convolution/Add" type="Add" version="opset1">
+		<layer id="9" name="__module.conv2/aten::_convolution/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -236,7 +156,7 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="14" name="aten::gelu/Gelu_1" type="Gelu" version="opset7">
+		<layer id="10" name="aten::gelu/Gelu_1" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -253,15 +173,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="15" name="Constant_120" type="Const" version="opset1">
-			<data element_type="i64" shape="3" offset="1820672" size="24" />
+		<layer id="11" name="Constant_120" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="3641344" size="24" />
 			<output>
 				<port id="0" precision="I64" names="28">
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="16" name="aten::permute/Transpose" type="Transpose" version="opset1">
+		<layer id="12" name="aten::permute/Transpose" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -280,37 +200,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="17" name="Constant_5511_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1500, 512" offset="1820696" size="1536000" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1500</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="18" name="Constant_5511" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1500</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="13" name="Constant_5893" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1500, 512" offset="3641368" size="3072000" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="19" name="aten::add/Add" type="Add" version="opset1">
+		<layer id="14" name="aten::add/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -332,15 +232,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="20" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="15" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="21" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="16" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -360,37 +260,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="22" name="Constant_5512_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="3356700" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="23" name="Constant_5512" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="17" name="Constant_5894" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="6713372" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="24" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="18" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -412,37 +292,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="25" name="Constant_5513_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="3357724" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="26" name="Constant_5513" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="19" name="Constant_5895" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="6715420" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="27" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="20" name="__module.layers.0.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -457,41 +317,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="90,hidden_states.1">
+				<port id="2" precision="FP32" names="88,hidden_states.1">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="28" name="Constant_5432_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="3358748" size="524288" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="29" name="Constant_5432" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="21" name="self.layers.0.self_attn.q_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="6717468" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.0.self_attn.q_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="30" name="Multiply_5407" type="MatMul" version="opset1">
+		<layer id="22" name="__module.layers.0.self_attn.q_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -512,37 +354,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="31" name="Constant_5514_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="3883036" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="32" name="Constant_5514" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="23" name="Constant_5896" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="7766044" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="33" name="__module.layers.0.self_attn/aten::mul/Multiply" type="Add" version="opset1">
+		<layer id="24" name="__module.layers.0.self_attn.q_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -557,25 +379,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="100,tensor.1">
+				<port id="2" precision="FP32" names="97,tensor.1">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="34" name="Constant_5674" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="25" name="Constant_6056" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="35" name="__module.layers.0.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
+		<layer id="26" name="__module.layers.0.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -588,7 +407,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="102">
+				<port id="2" precision="FP32" names="99">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -596,15 +415,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="36" name="__module.layers.0.self_attn/aten::transpose/Constant" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="27" name="__module.layers.0.self_attn/aten::transpose/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="37" name="__module.layers.0.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
+		<layer id="28" name="__module.layers.0.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -617,7 +436,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="103">
+				<port id="2" precision="FP32" names="100">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -625,34 +444,16 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="38" name="self.layers.0.self_attn.k_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="3884108" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.0.self_attn.k_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="39" name="self.layers.0.self_attn.k_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="29" name="self.layers.0.self_attn.k_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="7768140" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.0.self_attn.k_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="40" name="__module.layers.0.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="30" name="__module.layers.0.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -666,25 +467,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="106,tensor.3">
+				<port id="2" precision="FP32" names="103,tensor.3">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="41" name="Constant_5675" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="31" name="Constant_6057" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="42" name="__module.layers.0.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
+		<layer id="32" name="__module.layers.0.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -697,7 +495,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="108">
+				<port id="2" precision="FP32" names="105">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -705,15 +503,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="43" name="__module.layers.0.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="33" name="__module.layers.0.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="44" name="__module.layers.0.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
+		<layer id="34" name="__module.layers.0.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -726,86 +524,24 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="109">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="45" name="__module.layers.0.self_attn/aten::matmul/MatMul" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="true" />
-			<input>
-				<port id="0" precision="FP32">
+				<port id="2" precision="FP32" names="106">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
-				<port id="1" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-			</input>
-			<output>
-				<port id="2" precision="FP32" names="119,input.3">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="46" name="__module.layers.0.self_attn/aten::softmax/Softmax" type="SoftMax" version="opset8">
-			<data axis="-1" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32" names="120,input.5">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="47" name="self.layers.0.self_attn.v_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="4408396" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.0.self_attn.v_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
 			</output>
 		</layer>
-		<layer id="48" name="self.layers.0.self_attn.v_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="35" name="self.layers.0.self_attn.v_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="8816716" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.0.self_attn.v_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="49" name="__module.layers.0.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="36" name="__module.layers.0.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -826,37 +562,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="50" name="Constant_5515_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="4932684" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="51" name="Constant_5515" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="37" name="Constant_5897" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="9865292" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="52" name="__module.layers.0.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="38" name="__module.layers.0.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -871,25 +587,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="113,tensor.5">
+				<port id="2" precision="FP32" names="110,tensor.5">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="53" name="Constant_5676" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="39" name="Constant_6058" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="54" name="__module.layers.0.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
+		<layer id="40" name="__module.layers.0.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -902,7 +615,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="115">
+				<port id="2" precision="FP32" names="112">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -910,15 +623,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="55" name="__module.layers.0.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="41" name="__module.layers.0.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="56" name="__module.layers.0.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
+		<layer id="42" name="__module.layers.0.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -931,7 +644,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="116">
+				<port id="2" precision="FP32" names="113">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -939,14 +652,14 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="57" name="__module.layers.0.self_attn/aten::matmul/MatMul_1" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="false" />
+		<layer id="43" name="__module.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>8</dim>
 					<dim>1500</dim>
-					<dim>1500</dim>
+					<dim>64</dim>
 				</port>
 				<port id="1" precision="FP32">
 					<dim>-1</dim>
@@ -954,9 +667,15 @@
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>8</dim>
+					<dim>1500</dim>
+					<dim>64</dim>
+				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="122,attn_output.1">
+				<port id="3" precision="FP32" names="115,attn_output.1">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -964,15 +683,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="58" name="__module.layers.0.self_attn/aten::transpose/Constant_4" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="44" name="__module.layers.0.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="59" name="__module.layers.0.self_attn/aten::transpose/Transpose_4" type="Transpose" version="opset1">
+		<layer id="45" name="__module.layers.0.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -985,7 +704,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="123,attn_output.3">
+				<port id="2" precision="FP32" names="116,attn_output.3">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -993,18 +712,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="60" name="Constant_5677" type="Const" version="opset1">
-			<data element_type="i64" shape="3" offset="4933708" size="24" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="46" name="Constant_6059" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="9867340" size="24" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="61" name="__module.layers.0.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
+		<layer id="47" name="__module.layers.0.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1018,41 +734,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="125">
+				<port id="2" precision="FP32" names="118">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="62" name="self.layers.0.self_attn.out_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="4933732" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.0.self_attn.out_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="63" name="self.layers.0.self_attn.out_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="48" name="self.layers.0.self_attn.out_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="9867364" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.0.self_attn.out_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="64" name="__module.layers.0.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="49" name="__module.layers.0.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1073,37 +771,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="65" name="Constant_5516_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="5458020" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="66" name="Constant_5516" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="50" name="Constant_5898" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="10915940" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="67" name="__module.layers.0.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="51" name="__module.layers.0.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1118,14 +796,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="128,input.7">
+				<port id="2" precision="FP32" names="121,input.3">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="68" name="__module.layers.0/aten::add/Add" type="Add" version="opset1">
+		<layer id="52" name="__module.layers.0/aten::add/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1140,22 +818,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="130,residual.3">
+				<port id="2" precision="FP32" names="123,residual.3">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="69" name="__module.layers.0.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="53" name="__module.layers.0.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="70" name="__module.layers.0.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="54" name="__module.layers.0.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1175,42 +853,22 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="71" name="Constant_5517_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="5459044" size="1024" />
+		<layer id="55" name="Constant_5899" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="10917988" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="72" name="Constant_5517" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
+		<layer id="56" name="__module.layers.0.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
 			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="73" name="__module.layers.0.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
-			<data auto_broadcast="numpy" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>1500</dim>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 				<port id="1" precision="FP32">
@@ -1227,37 +885,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="74" name="Constant_5518_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="5460068" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="75" name="Constant_5518" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="57" name="Constant_5900" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="10920036" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="76" name="__module.layers.0.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="58" name="__module.layers.0.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1272,41 +910,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="134">
+				<port id="2" precision="FP32" names="127">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="77" name="self.layers.0.fc1.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="2048, 512" offset="5461092" size="2097152" />
+		<layer id="59" name="self.layers.0.fc1.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="2048, 512" offset="10922084" size="4194304" />
 			<output>
-				<port id="0" precision="FP16" names="self.layers.0.fc1.weight">
+				<port id="0" precision="FP32" names="self.layers.0.fc1.weight">
 					<dim>2048</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="78" name="self.layers.0.fc1.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="79" name="__module.layers.0.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="60" name="__module.layers.0.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1327,37 +947,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="80" name="Constant_5519_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 2048" offset="7558244" size="4096" />
+		<layer id="61" name="Constant_5901" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 2048" offset="15116388" size="8192" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="81" name="Constant_5519" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="82" name="__module.layers.0.fc1/aten::linear/Add" type="Add" version="opset1">
+		<layer id="62" name="__module.layers.0.fc1/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1372,14 +972,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="137">
+				<port id="2" precision="FP32" names="130">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="83" name="__module.layers.0.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
+		<layer id="63" name="__module.layers.0.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1389,41 +989,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="1" precision="FP32" names="138,input.9">
+				<port id="1" precision="FP32" names="131,input.5">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="84" name="self.layers.0.fc2.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 2048" offset="7562340" size="2097152" />
+		<layer id="64" name="self.layers.0.fc2.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 2048" offset="15124580" size="4194304" />
 			<output>
-				<port id="0" precision="FP16" names="self.layers.0.fc2.weight">
+				<port id="0" precision="FP32" names="self.layers.0.fc2.weight">
 					<dim>512</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="85" name="self.layers.0.fc2.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="86" name="__module.layers.0.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="65" name="__module.layers.0.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1444,37 +1026,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="87" name="Constant_5520_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="9659492" size="1024" />
+		<layer id="66" name="Constant_5902" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="19318884" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="88" name="Constant_5520" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="89" name="__module.layers.0.fc2/aten::linear/Add" type="Add" version="opset1">
+		<layer id="67" name="__module.layers.0.fc2/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1489,14 +1051,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="142,input.11">
+				<port id="2" precision="FP32" names="135,input.7">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="90" name="__module.layers.0/aten::add/Add_1" type="Add" version="opset1">
+		<layer id="68" name="__module.layers.0/aten::add/Add_1" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1511,22 +1073,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="144,residual.5">
+				<port id="2" precision="FP32" names="137,residual.5">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="91" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="69" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="92" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="70" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1546,37 +1108,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="93" name="Constant_5521_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="9660516" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="94" name="Constant_5521" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="71" name="Constant_5903" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="19320932" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="95" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="72" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1598,37 +1140,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="96" name="Constant_5522_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="9661540" size="1024" />
+		<layer id="73" name="Constant_5904" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="19322980" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="97" name="Constant_5522" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="98" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="74" name="__module.layers.1.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1643,41 +1165,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="169,hidden_states.7">
+				<port id="2" precision="FP32" names="160,hidden_states.7">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="99" name="Constant_5435_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="9662564" size="524288" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="100" name="Constant_5435" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="75" name="self.layers.1.self_attn.q_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="19325028" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.1.self_attn.q_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="101" name="Multiply_5411" type="MatMul" version="opset1">
+		<layer id="76" name="__module.layers.1.self_attn.q_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1698,37 +1202,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="102" name="Constant_5523_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="10186852" size="1024" />
+		<layer id="77" name="Constant_5905" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="20373604" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="103" name="Constant_5523" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="104" name="__module.layers.1.self_attn/aten::mul/Multiply" type="Add" version="opset1">
+		<layer id="78" name="__module.layers.1.self_attn.q_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1743,25 +1227,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="179,tensor.7">
+				<port id="2" precision="FP32" names="169,tensor.7">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="105" name="Constant_5678" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="79" name="Constant_6060" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="106" name="__module.layers.1.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
+		<layer id="80" name="__module.layers.1.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1774,7 +1255,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="181">
+				<port id="2" precision="FP32" names="171">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -1782,15 +1263,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="107" name="__module.layers.1.self_attn/aten::transpose/Constant" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="81" name="__module.layers.1.self_attn/aten::transpose/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="108" name="__module.layers.1.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
+		<layer id="82" name="__module.layers.1.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -1803,7 +1284,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="182">
+				<port id="2" precision="FP32" names="172">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -1811,34 +1292,16 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="109" name="self.layers.1.self_attn.k_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="10187876" size="524288" />
+		<layer id="83" name="self.layers.1.self_attn.k_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="20375652" size="1048576" />
 			<output>
-				<port id="0" precision="FP16" names="self.layers.1.self_attn.k_proj.weight">
+				<port id="0" precision="FP32" names="self.layers.1.self_attn.k_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="110" name="self.layers.1.self_attn.k_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="111" name="__module.layers.1.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="84" name="__module.layers.1.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1852,25 +1315,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="185,tensor.9">
+				<port id="2" precision="FP32" names="175,tensor.9">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="112" name="Constant_5679" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="85" name="Constant_6061" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="113" name="__module.layers.1.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
+		<layer id="86" name="__module.layers.1.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -1883,7 +1343,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="187">
+				<port id="2" precision="FP32" names="177">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -1891,15 +1351,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="114" name="__module.layers.1.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="87" name="__module.layers.1.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="115" name="__module.layers.1.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
+		<layer id="88" name="__module.layers.1.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -1912,7 +1372,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="188">
+				<port id="2" precision="FP32" names="178">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -1920,78 +1380,16 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="116" name="__module.layers.1.self_attn/aten::matmul/MatMul" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="true" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-				<port id="1" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-			</input>
-			<output>
-				<port id="2" precision="FP32" names="198,input.13">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="117" name="__module.layers.1.self_attn/aten::softmax/Softmax" type="SoftMax" version="opset8">
-			<data axis="-1" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</input>
+		<layer id="89" name="self.layers.1.self_attn.v_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="21424228" size="1048576" />
 			<output>
-				<port id="1" precision="FP32" names="199,input.15">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="118" name="self.layers.1.self_attn.v_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="10712164" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.1.self_attn.v_proj.weight">
+				<port id="0" precision="FP32" names="self.layers.1.self_attn.v_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="119" name="self.layers.1.self_attn.v_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="120" name="__module.layers.1.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="90" name="__module.layers.1.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2012,37 +1410,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="121" name="Constant_5524_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="11236452" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="122" name="Constant_5524" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="91" name="Constant_5906" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="22472804" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="123" name="__module.layers.1.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="92" name="__module.layers.1.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2057,25 +1435,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="192,tensor.11">
+				<port id="2" precision="FP32" names="182,tensor.11">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="124" name="Constant_5680" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="93" name="Constant_6062" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="125" name="__module.layers.1.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
+		<layer id="94" name="__module.layers.1.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2088,7 +1463,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="194">
+				<port id="2" precision="FP32" names="184">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -2096,15 +1471,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="126" name="__module.layers.1.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="95" name="__module.layers.1.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="127" name="__module.layers.1.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
+		<layer id="96" name="__module.layers.1.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -2117,7 +1492,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="195">
+				<port id="2" precision="FP32" names="185">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -2125,14 +1500,14 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="128" name="__module.layers.1.self_attn/aten::matmul/MatMul_1" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="false" />
+		<layer id="97" name="__module.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>8</dim>
 					<dim>1500</dim>
-					<dim>1500</dim>
+					<dim>64</dim>
 				</port>
 				<port id="1" precision="FP32">
 					<dim>-1</dim>
@@ -2140,9 +1515,15 @@
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>8</dim>
+					<dim>1500</dim>
+					<dim>64</dim>
+				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="201,attn_output.5">
+				<port id="3" precision="FP32" names="187,attn_output.5">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -2150,15 +1531,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="129" name="__module.layers.1.self_attn/aten::transpose/Constant_4" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="98" name="__module.layers.1.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="130" name="__module.layers.1.self_attn/aten::transpose/Transpose_4" type="Transpose" version="opset1">
+		<layer id="99" name="__module.layers.1.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -2171,7 +1552,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="202,attn_output.7">
+				<port id="2" precision="FP32" names="188,attn_output.7">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -2179,18 +1560,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="131" name="Constant_5681" type="Const" version="opset1">
-			<data element_type="i64" shape="3" offset="4933708" size="24" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="100" name="Constant_6063" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="9867340" size="24" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="132" name="__module.layers.1.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
+		<layer id="101" name="__module.layers.1.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2204,41 +1582,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="204">
+				<port id="2" precision="FP32" names="190">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="133" name="self.layers.1.self_attn.out_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="11237476" size="524288" />
+		<layer id="102" name="self.layers.1.self_attn.out_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="22474852" size="1048576" />
 			<output>
-				<port id="0" precision="FP16" names="self.layers.1.self_attn.out_proj.weight">
+				<port id="0" precision="FP32" names="self.layers.1.self_attn.out_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="134" name="self.layers.1.self_attn.out_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="135" name="__module.layers.1.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="103" name="__module.layers.1.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2259,37 +1619,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="136" name="Constant_5525_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="11761764" size="1024" />
+		<layer id="104" name="Constant_5907" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="23523428" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="137" name="Constant_5525" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="138" name="__module.layers.1.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="105" name="__module.layers.1.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2304,14 +1644,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="207,input.17">
+				<port id="2" precision="FP32" names="193,input.9">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="139" name="__module.layers.1/aten::add/Add" type="Add" version="opset1">
+		<layer id="106" name="__module.layers.1/aten::add/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2326,22 +1666,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="209,residual.7">
+				<port id="2" precision="FP32" names="195,residual.7">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="140" name="__module.layers.1.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="107" name="__module.layers.1.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="141" name="__module.layers.1.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="108" name="__module.layers.1.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2361,37 +1701,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="142" name="Constant_5526_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="11762788" size="1024" />
+		<layer id="109" name="Constant_5908" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="23525476" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="143" name="Constant_5526" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="144" name="__module.layers.1.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="110" name="__module.layers.1.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2413,37 +1733,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="145" name="Constant_5527_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="11763812" size="1024" />
+		<layer id="111" name="Constant_5909" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="23527524" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="146" name="Constant_5527" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="147" name="__module.layers.1.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="112" name="__module.layers.1.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2458,41 +1758,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="213">
+				<port id="2" precision="FP32" names="199">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="148" name="self.layers.1.fc1.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="2048, 512" offset="11764836" size="2097152" />
+		<layer id="113" name="self.layers.1.fc1.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="2048, 512" offset="23529572" size="4194304" />
 			<output>
-				<port id="0" precision="FP16" names="self.layers.1.fc1.weight">
+				<port id="0" precision="FP32" names="self.layers.1.fc1.weight">
 					<dim>2048</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="149" name="self.layers.1.fc1.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="150" name="__module.layers.1.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="114" name="__module.layers.1.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2513,37 +1795,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="151" name="Constant_5528_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 2048" offset="13861988" size="4096" />
+		<layer id="115" name="Constant_5910" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 2048" offset="27723876" size="8192" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="152" name="Constant_5528" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="153" name="__module.layers.1.fc1/aten::linear/Add" type="Add" version="opset1">
+		<layer id="116" name="__module.layers.1.fc1/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2558,14 +1820,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="216">
+				<port id="2" precision="FP32" names="202">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="154" name="__module.layers.1.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
+		<layer id="117" name="__module.layers.1.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2575,41 +1837,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="1" precision="FP32" names="217,input.19">
+				<port id="1" precision="FP32" names="203,input.11">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="155" name="self.layers.1.fc2.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 2048" offset="13866084" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.1.fc2.weight">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="156" name="self.layers.1.fc2.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="118" name="self.layers.1.fc2.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 2048" offset="27732068" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.1.fc2.weight">
 					<dim>512</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="157" name="__module.layers.1.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="119" name="__module.layers.1.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2630,37 +1874,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="158" name="Constant_5529_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="15963236" size="1024" />
+		<layer id="120" name="Constant_5911" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="31926372" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="159" name="Constant_5529" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="160" name="__module.layers.1.fc2/aten::linear/Add" type="Add" version="opset1">
+		<layer id="121" name="__module.layers.1.fc2/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2675,14 +1899,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="221,input.21">
+				<port id="2" precision="FP32" names="207,input.13">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="161" name="__module.layers.1/aten::add/Add_1" type="Add" version="opset1">
+		<layer id="122" name="__module.layers.1/aten::add/Add_1" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2697,22 +1921,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="223,residual.9">
+				<port id="2" precision="FP32" names="209,residual.9">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="162" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="123" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="163" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="124" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2732,37 +1956,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="164" name="Constant_5530_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="15964260" size="1024" />
+		<layer id="125" name="Constant_5912" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="31928420" size="2048" />
 			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="165" name="Constant_5530" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="166" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="126" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2784,37 +1988,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="167" name="Constant_5531_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="15965284" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="168" name="Constant_5531" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="127" name="Constant_5913" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="31930468" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="169" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="128" name="__module.layers.2.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2829,41 +2013,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="248,hidden_states.13">
+				<port id="2" precision="FP32" names="232,hidden_states.13">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="170" name="Constant_5438_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="15966308" size="524288" />
+		<layer id="129" name="self.layers.2.self_attn.q_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="31932516" size="1048576" />
 			<output>
-				<port id="0" precision="FP16">
+				<port id="0" precision="FP32" names="self.layers.2.self_attn.q_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="171" name="Constant_5438" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="172" name="Multiply_5415" type="MatMul" version="opset1">
+		<layer id="130" name="__module.layers.2.self_attn.q_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2884,37 +2050,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="173" name="Constant_5532_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="16490596" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="174" name="Constant_5532" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="131" name="Constant_5914" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="32981092" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="175" name="__module.layers.2.self_attn/aten::mul/Multiply" type="Add" version="opset1">
+		<layer id="132" name="__module.layers.2.self_attn.q_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2929,25 +2075,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="258,tensor.13">
+				<port id="2" precision="FP32" names="241,tensor.13">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="176" name="Constant_5682" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="133" name="Constant_6064" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="177" name="__module.layers.2.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
+		<layer id="134" name="__module.layers.2.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -2960,7 +2103,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="260">
+				<port id="2" precision="FP32" names="243">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -2968,15 +2111,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="178" name="__module.layers.2.self_attn/aten::transpose/Constant" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="135" name="__module.layers.2.self_attn/aten::transpose/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="179" name="__module.layers.2.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
+		<layer id="136" name="__module.layers.2.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -2989,7 +2132,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="261">
+				<port id="2" precision="FP32" names="244">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -2997,34 +2140,16 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="180" name="self.layers.2.self_attn.k_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="16491620" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.2.self_attn.k_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="181" name="self.layers.2.self_attn.k_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="137" name="self.layers.2.self_attn.k_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="32983140" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.2.self_attn.k_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="182" name="__module.layers.2.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="138" name="__module.layers.2.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3038,25 +2163,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="264,tensor.15">
+				<port id="2" precision="FP32" names="247,tensor.15">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="183" name="Constant_5683" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="139" name="Constant_6065" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="184" name="__module.layers.2.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
+		<layer id="140" name="__module.layers.2.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3069,7 +2191,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="266">
+				<port id="2" precision="FP32" names="249">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -3077,15 +2199,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="185" name="__module.layers.2.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="141" name="__module.layers.2.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="186" name="__module.layers.2.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
+		<layer id="142" name="__module.layers.2.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -3098,86 +2220,24 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="267">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="187" name="__module.layers.2.self_attn/aten::matmul/MatMul" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="true" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-				<port id="1" precision="FP32">
+				<port id="2" precision="FP32" names="250">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
-			</input>
-			<output>
-				<port id="2" precision="FP32" names="277,input.23">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="188" name="__module.layers.2.self_attn/aten::softmax/Softmax" type="SoftMax" version="opset8">
-			<data axis="-1" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32" names="278,input.25">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="189" name="self.layers.2.self_attn.v_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="17015908" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.2.self_attn.v_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
 			</output>
 		</layer>
-		<layer id="190" name="self.layers.2.self_attn.v_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="143" name="self.layers.2.self_attn.v_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="34031716" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.2.self_attn.v_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="191" name="__module.layers.2.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="144" name="__module.layers.2.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3198,37 +2258,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="192" name="Constant_5533_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="17540196" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="193" name="Constant_5533" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="145" name="Constant_5915" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="35080292" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="194" name="__module.layers.2.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="146" name="__module.layers.2.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3243,25 +2283,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="271,tensor.17">
+				<port id="2" precision="FP32" names="254,tensor.17">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="195" name="Constant_5684" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="147" name="Constant_6066" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="196" name="__module.layers.2.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
+		<layer id="148" name="__module.layers.2.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3274,7 +2311,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="273">
+				<port id="2" precision="FP32" names="256">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -3282,15 +2319,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="197" name="__module.layers.2.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="149" name="__module.layers.2.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="198" name="__module.layers.2.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
+		<layer id="150" name="__module.layers.2.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -3303,7 +2340,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="274">
+				<port id="2" precision="FP32" names="257">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -3311,14 +2348,14 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="199" name="__module.layers.2.self_attn/aten::matmul/MatMul_1" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="false" />
+		<layer id="151" name="__module.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>8</dim>
 					<dim>1500</dim>
-					<dim>1500</dim>
+					<dim>64</dim>
 				</port>
 				<port id="1" precision="FP32">
 					<dim>-1</dim>
@@ -3326,9 +2363,15 @@
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>8</dim>
+					<dim>1500</dim>
+					<dim>64</dim>
+				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="280,attn_output.9">
+				<port id="3" precision="FP32" names="259,attn_output.9">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -3336,15 +2379,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="200" name="__module.layers.2.self_attn/aten::transpose/Constant_4" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="152" name="__module.layers.2.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="201" name="__module.layers.2.self_attn/aten::transpose/Transpose_4" type="Transpose" version="opset1">
+		<layer id="153" name="__module.layers.2.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -3357,7 +2400,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="281,attn_output.11">
+				<port id="2" precision="FP32" names="260,attn_output.11">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -3365,18 +2408,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="202" name="Constant_5685" type="Const" version="opset1">
-			<data element_type="i64" shape="3" offset="4933708" size="24" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="154" name="Constant_6067" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="9867340" size="24" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="203" name="__module.layers.2.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
+		<layer id="155" name="__module.layers.2.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3390,41 +2430,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="283">
+				<port id="2" precision="FP32" names="262">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="204" name="self.layers.2.self_attn.out_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="17541220" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.2.self_attn.out_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="205" name="self.layers.2.self_attn.out_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="156" name="self.layers.2.self_attn.out_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="35082340" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.2.self_attn.out_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="206" name="__module.layers.2.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="157" name="__module.layers.2.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3445,37 +2467,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="207" name="Constant_5534_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="18065508" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="208" name="Constant_5534" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="158" name="Constant_5916" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="36130916" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="209" name="__module.layers.2.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="159" name="__module.layers.2.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3490,14 +2492,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="286,input.27">
+				<port id="2" precision="FP32" names="265,input.15">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="210" name="__module.layers.2/aten::add/Add" type="Add" version="opset1">
+		<layer id="160" name="__module.layers.2/aten::add/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3512,22 +2514,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="288,residual.11">
+				<port id="2" precision="FP32" names="267,residual.11">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="211" name="__module.layers.2.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="161" name="__module.layers.2.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="212" name="__module.layers.2.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="162" name="__module.layers.2.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3547,37 +2549,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="213" name="Constant_5535_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="18066532" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="214" name="Constant_5535" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="163" name="Constant_5917" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="36132964" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="215" name="__module.layers.2.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="164" name="__module.layers.2.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3599,37 +2581,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="216" name="Constant_5536_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="18067556" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="217" name="Constant_5536" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="165" name="Constant_5918" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="36135012" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="218" name="__module.layers.2.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="166" name="__module.layers.2.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3644,41 +2606,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="292">
+				<port id="2" precision="FP32" names="271">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="219" name="self.layers.2.fc1.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="2048, 512" offset="18068580" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.2.fc1.weight">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="220" name="self.layers.2.fc1.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="167" name="self.layers.2.fc1.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="2048, 512" offset="36137060" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.2.fc1.weight">
 					<dim>2048</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="221" name="__module.layers.2.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="168" name="__module.layers.2.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3699,37 +2643,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="222" name="Constant_5537_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 2048" offset="20165732" size="4096" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="223" name="Constant_5537" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="169" name="Constant_5919" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 2048" offset="40331364" size="8192" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="224" name="__module.layers.2.fc1/aten::linear/Add" type="Add" version="opset1">
+		<layer id="170" name="__module.layers.2.fc1/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3744,14 +2668,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="295">
+				<port id="2" precision="FP32" names="274">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="225" name="__module.layers.2.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
+		<layer id="171" name="__module.layers.2.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3761,41 +2685,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="1" precision="FP32" names="296,input.29">
+				<port id="1" precision="FP32" names="275,input.17">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="226" name="self.layers.2.fc2.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 2048" offset="20169828" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.2.fc2.weight">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="227" name="self.layers.2.fc2.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="172" name="self.layers.2.fc2.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 2048" offset="40339556" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.2.fc2.weight">
 					<dim>512</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="228" name="__module.layers.2.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="173" name="__module.layers.2.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3816,37 +2722,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="229" name="Constant_5538_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="22266980" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="230" name="Constant_5538" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="174" name="Constant_5920" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="44533860" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="231" name="__module.layers.2.fc2/aten::linear/Add" type="Add" version="opset1">
+		<layer id="175" name="__module.layers.2.fc2/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3861,14 +2747,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="300,input.31">
+				<port id="2" precision="FP32" names="279,input.19">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="232" name="__module.layers.2/aten::add/Add_1" type="Add" version="opset1">
+		<layer id="176" name="__module.layers.2/aten::add/Add_1" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3883,22 +2769,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="302,residual.13">
+				<port id="2" precision="FP32" names="281,residual.13">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="233" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="177" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="234" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="178" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3918,37 +2804,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="235" name="Constant_5539_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="22268004" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="236" name="Constant_5539" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="179" name="Constant_5921" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="44535908" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="237" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="180" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -3970,37 +2836,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="238" name="Constant_5540_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="22269028" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="239" name="Constant_5540" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="181" name="Constant_5922" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="44537956" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="240" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="182" name="__module.layers.3.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4015,41 +2861,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="327,hidden_states.19">
+				<port id="2" precision="FP32" names="304,hidden_states.19">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="241" name="Constant_5441_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="22270052" size="524288" />
+		<layer id="183" name="self.layers.3.self_attn.q_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="44540004" size="1048576" />
 			<output>
-				<port id="0" precision="FP16">
+				<port id="0" precision="FP32" names="self.layers.3.self_attn.q_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="242" name="Constant_5441" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="243" name="Multiply_5419" type="MatMul" version="opset1">
+		<layer id="184" name="__module.layers.3.self_attn.q_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4070,37 +2898,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="244" name="Constant_5541_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="22794340" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="245" name="Constant_5541" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="185" name="Constant_5923" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="45588580" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="246" name="__module.layers.3.self_attn/aten::mul/Multiply" type="Add" version="opset1">
+		<layer id="186" name="__module.layers.3.self_attn.q_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4115,25 +2923,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="337,tensor.19">
+				<port id="2" precision="FP32" names="313,tensor.19">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="247" name="Constant_5686" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="187" name="Constant_6068" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="248" name="__module.layers.3.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
+		<layer id="188" name="__module.layers.3.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4146,7 +2951,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="339">
+				<port id="2" precision="FP32" names="315">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -4154,15 +2959,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="249" name="__module.layers.3.self_attn/aten::transpose/Constant" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="189" name="__module.layers.3.self_attn/aten::transpose/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="250" name="__module.layers.3.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
+		<layer id="190" name="__module.layers.3.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -4175,7 +2980,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="340">
+				<port id="2" precision="FP32" names="316">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -4183,34 +2988,16 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="251" name="self.layers.3.self_attn.k_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="22795364" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.3.self_attn.k_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="252" name="self.layers.3.self_attn.k_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="191" name="self.layers.3.self_attn.k_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="45590628" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.3.self_attn.k_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="253" name="__module.layers.3.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="192" name="__module.layers.3.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4224,25 +3011,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="343,tensor.21">
+				<port id="2" precision="FP32" names="319,tensor.21">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="254" name="Constant_5687" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="193" name="Constant_6069" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="255" name="__module.layers.3.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
+		<layer id="194" name="__module.layers.3.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4255,7 +3039,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="345">
+				<port id="2" precision="FP32" names="321">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -4263,15 +3047,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="256" name="__module.layers.3.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="195" name="__module.layers.3.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="257" name="__module.layers.3.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
+		<layer id="196" name="__module.layers.3.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -4284,86 +3068,24 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="346">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="258" name="__module.layers.3.self_attn/aten::matmul/MatMul" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="true" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-				<port id="1" precision="FP32">
+				<port id="2" precision="FP32" names="322">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
-			</input>
-			<output>
-				<port id="2" precision="FP32" names="356,input.33">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="259" name="__module.layers.3.self_attn/aten::softmax/Softmax" type="SoftMax" version="opset8">
-			<data axis="-1" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32" names="357,input.35">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="260" name="self.layers.3.self_attn.v_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="23319652" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.3.self_attn.v_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
 			</output>
 		</layer>
-		<layer id="261" name="self.layers.3.self_attn.v_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="197" name="self.layers.3.self_attn.v_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="46639204" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.3.self_attn.v_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="262" name="__module.layers.3.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="198" name="__module.layers.3.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4384,37 +3106,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="263" name="Constant_5542_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="23843940" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="264" name="Constant_5542" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="199" name="Constant_5924" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="47687780" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="265" name="__module.layers.3.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="200" name="__module.layers.3.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4429,25 +3131,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="350,tensor.23">
+				<port id="2" precision="FP32" names="326,tensor.23">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="266" name="Constant_5688" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="201" name="Constant_6070" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="267" name="__module.layers.3.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
+		<layer id="202" name="__module.layers.3.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4460,7 +3159,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="352">
+				<port id="2" precision="FP32" names="328">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -4468,15 +3167,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="268" name="__module.layers.3.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="203" name="__module.layers.3.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="269" name="__module.layers.3.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
+		<layer id="204" name="__module.layers.3.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -4489,7 +3188,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="353">
+				<port id="2" precision="FP32" names="329">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -4497,14 +3196,14 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="270" name="__module.layers.3.self_attn/aten::matmul/MatMul_1" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="false" />
+		<layer id="205" name="__module.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>8</dim>
 					<dim>1500</dim>
-					<dim>1500</dim>
+					<dim>64</dim>
 				</port>
 				<port id="1" precision="FP32">
 					<dim>-1</dim>
@@ -4512,9 +3211,15 @@
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>8</dim>
+					<dim>1500</dim>
+					<dim>64</dim>
+				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="359,attn_output.13">
+				<port id="3" precision="FP32" names="331,attn_output.13">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -4522,15 +3227,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="271" name="__module.layers.3.self_attn/aten::transpose/Constant_4" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="206" name="__module.layers.3.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="272" name="__module.layers.3.self_attn/aten::transpose/Transpose_4" type="Transpose" version="opset1">
+		<layer id="207" name="__module.layers.3.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -4543,7 +3248,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="360,attn_output.15">
+				<port id="2" precision="FP32" names="332,attn_output.15">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -4551,18 +3256,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="273" name="Constant_5689" type="Const" version="opset1">
-			<data element_type="i64" shape="3" offset="4933708" size="24" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="208" name="Constant_6071" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="9867340" size="24" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="274" name="__module.layers.3.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
+		<layer id="209" name="__module.layers.3.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4576,41 +3278,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="362">
+				<port id="2" precision="FP32" names="334">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="275" name="self.layers.3.self_attn.out_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="23844964" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.3.self_attn.out_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="276" name="self.layers.3.self_attn.out_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="210" name="self.layers.3.self_attn.out_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="47689828" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.3.self_attn.out_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="277" name="__module.layers.3.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="211" name="__module.layers.3.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4631,37 +3315,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="278" name="Constant_5543_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="24369252" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="279" name="Constant_5543" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="212" name="Constant_5925" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="48738404" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="280" name="__module.layers.3.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="213" name="__module.layers.3.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4676,14 +3340,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="365,input.37">
+				<port id="2" precision="FP32" names="337,input.21">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="281" name="__module.layers.3/aten::add/Add" type="Add" version="opset1">
+		<layer id="214" name="__module.layers.3/aten::add/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4698,22 +3362,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="367,residual.15">
+				<port id="2" precision="FP32" names="339,residual.15">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="282" name="__module.layers.3.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="215" name="__module.layers.3.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="283" name="__module.layers.3.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="216" name="__module.layers.3.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4733,37 +3397,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="284" name="Constant_5544_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="24370276" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="285" name="Constant_5544" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="217" name="Constant_5926" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="48740452" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="286" name="__module.layers.3.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="218" name="__module.layers.3.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4785,37 +3429,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="287" name="Constant_5545_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="24371300" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="288" name="Constant_5545" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="219" name="Constant_5927" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="48742500" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="289" name="__module.layers.3.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="220" name="__module.layers.3.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4830,41 +3454,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="371">
+				<port id="2" precision="FP32" names="343">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="290" name="self.layers.3.fc1.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="2048, 512" offset="24372324" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.3.fc1.weight">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="291" name="self.layers.3.fc1.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="221" name="self.layers.3.fc1.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="2048, 512" offset="48744548" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.3.fc1.weight">
 					<dim>2048</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="292" name="__module.layers.3.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="222" name="__module.layers.3.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4885,37 +3491,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="293" name="Constant_5546_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 2048" offset="26469476" size="4096" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="294" name="Constant_5546" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="223" name="Constant_5928" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 2048" offset="52938852" size="8192" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="295" name="__module.layers.3.fc1/aten::linear/Add" type="Add" version="opset1">
+		<layer id="224" name="__module.layers.3.fc1/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4930,14 +3516,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="374">
+				<port id="2" precision="FP32" names="346">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="296" name="__module.layers.3.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
+		<layer id="225" name="__module.layers.3.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -4947,41 +3533,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="1" precision="FP32" names="375,input.39">
+				<port id="1" precision="FP32" names="347,input.23">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="297" name="self.layers.3.fc2.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 2048" offset="26473572" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.3.fc2.weight">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="298" name="self.layers.3.fc2.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="226" name="self.layers.3.fc2.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 2048" offset="52947044" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.3.fc2.weight">
 					<dim>512</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="299" name="__module.layers.3.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="227" name="__module.layers.3.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5002,37 +3570,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="300" name="Constant_5547_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="28570724" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="301" name="Constant_5547" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="228" name="Constant_5929" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="57141348" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="302" name="__module.layers.3.fc2/aten::linear/Add" type="Add" version="opset1">
+		<layer id="229" name="__module.layers.3.fc2/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5047,14 +3595,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="379,input.41">
+				<port id="2" precision="FP32" names="351,input.25">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="303" name="__module.layers.3/aten::add/Add_1" type="Add" version="opset1">
+		<layer id="230" name="__module.layers.3/aten::add/Add_1" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5069,22 +3617,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="381,residual.17">
+				<port id="2" precision="FP32" names="353,residual.17">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="304" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="231" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="305" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="232" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5104,37 +3652,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="306" name="Constant_5548_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="28571748" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="307" name="Constant_5548" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="233" name="Constant_5930" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="57143396" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="308" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="234" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5156,37 +3684,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="309" name="Constant_5549_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="28572772" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="310" name="Constant_5549" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="235" name="Constant_5931" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="57145444" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="311" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="236" name="__module.layers.4.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5201,41 +3709,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="406,hidden_states.25">
+				<port id="2" precision="FP32" names="376,hidden_states.25">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="312" name="Constant_5444_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="28573796" size="524288" />
+		<layer id="237" name="self.layers.4.self_attn.q_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="57147492" size="1048576" />
 			<output>
-				<port id="0" precision="FP16">
+				<port id="0" precision="FP32" names="self.layers.4.self_attn.q_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="313" name="Constant_5444" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="314" name="Multiply_5423" type="MatMul" version="opset1">
+		<layer id="238" name="__module.layers.4.self_attn.q_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5256,37 +3746,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="315" name="Constant_5550_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="29098084" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="316" name="Constant_5550" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="239" name="Constant_5932" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="58196068" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="317" name="__module.layers.4.self_attn/aten::mul/Multiply" type="Add" version="opset1">
+		<layer id="240" name="__module.layers.4.self_attn.q_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5301,25 +3771,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="416,tensor.25">
+				<port id="2" precision="FP32" names="385,tensor.25">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="318" name="Constant_5690" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="241" name="Constant_6072" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="319" name="__module.layers.4.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
+		<layer id="242" name="__module.layers.4.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5332,7 +3799,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="418">
+				<port id="2" precision="FP32" names="387">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -5340,15 +3807,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="320" name="__module.layers.4.self_attn/aten::transpose/Constant" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="243" name="__module.layers.4.self_attn/aten::transpose/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="321" name="__module.layers.4.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
+		<layer id="244" name="__module.layers.4.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -5361,7 +3828,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="419">
+				<port id="2" precision="FP32" names="388">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -5369,34 +3836,16 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="322" name="self.layers.4.self_attn.k_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="29099108" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.4.self_attn.k_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="323" name="self.layers.4.self_attn.k_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="245" name="self.layers.4.self_attn.k_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="58198116" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.4.self_attn.k_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="324" name="__module.layers.4.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="246" name="__module.layers.4.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5410,25 +3859,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="422,tensor.27">
+				<port id="2" precision="FP32" names="391,tensor.27">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="325" name="Constant_5691" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="247" name="Constant_6073" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="326" name="__module.layers.4.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
+		<layer id="248" name="__module.layers.4.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5441,7 +3887,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="424">
+				<port id="2" precision="FP32" names="393">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -5449,15 +3895,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="327" name="__module.layers.4.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="249" name="__module.layers.4.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="328" name="__module.layers.4.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
+		<layer id="250" name="__module.layers.4.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -5470,86 +3916,24 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="425">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="329" name="__module.layers.4.self_attn/aten::matmul/MatMul" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="true" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-				<port id="1" precision="FP32">
+				<port id="2" precision="FP32" names="394">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
-			</input>
-			<output>
-				<port id="2" precision="FP32" names="435,input.43">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="330" name="__module.layers.4.self_attn/aten::softmax/Softmax" type="SoftMax" version="opset8">
-			<data axis="-1" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32" names="436,input.45">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="331" name="self.layers.4.self_attn.v_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="29623396" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.4.self_attn.v_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
 			</output>
 		</layer>
-		<layer id="332" name="self.layers.4.self_attn.v_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="251" name="self.layers.4.self_attn.v_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="59246692" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.4.self_attn.v_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="333" name="__module.layers.4.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="252" name="__module.layers.4.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5570,37 +3954,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="334" name="Constant_5551_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="30147684" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="335" name="Constant_5551" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="253" name="Constant_5933" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="60295268" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="336" name="__module.layers.4.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="254" name="__module.layers.4.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5615,25 +3979,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="429,tensor.29">
+				<port id="2" precision="FP32" names="398,tensor.29">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="337" name="Constant_5692" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="255" name="Constant_6074" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="338" name="__module.layers.4.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
+		<layer id="256" name="__module.layers.4.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5646,7 +4007,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="431">
+				<port id="2" precision="FP32" names="400">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -5654,15 +4015,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="339" name="__module.layers.4.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="257" name="__module.layers.4.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="340" name="__module.layers.4.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
+		<layer id="258" name="__module.layers.4.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -5675,7 +4036,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="432">
+				<port id="2" precision="FP32" names="401">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -5683,14 +4044,14 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="341" name="__module.layers.4.self_attn/aten::matmul/MatMul_1" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="false" />
+		<layer id="259" name="__module.layers.4.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>8</dim>
 					<dim>1500</dim>
-					<dim>1500</dim>
+					<dim>64</dim>
 				</port>
 				<port id="1" precision="FP32">
 					<dim>-1</dim>
@@ -5698,9 +4059,15 @@
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>8</dim>
+					<dim>1500</dim>
+					<dim>64</dim>
+				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="438,attn_output.17">
+				<port id="3" precision="FP32" names="403,attn_output.17">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -5708,15 +4075,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="342" name="__module.layers.4.self_attn/aten::transpose/Constant_4" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="260" name="__module.layers.4.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="343" name="__module.layers.4.self_attn/aten::transpose/Transpose_4" type="Transpose" version="opset1">
+		<layer id="261" name="__module.layers.4.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -5729,7 +4096,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="439,attn_output.19">
+				<port id="2" precision="FP32" names="404,attn_output.19">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -5737,18 +4104,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="344" name="Constant_5693" type="Const" version="opset1">
-			<data element_type="i64" shape="3" offset="4933708" size="24" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="262" name="Constant_6075" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="9867340" size="24" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="345" name="__module.layers.4.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
+		<layer id="263" name="__module.layers.4.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5762,41 +4126,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="441">
+				<port id="2" precision="FP32" names="406">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="346" name="self.layers.4.self_attn.out_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="30148708" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.4.self_attn.out_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="347" name="self.layers.4.self_attn.out_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="264" name="self.layers.4.self_attn.out_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="60297316" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.4.self_attn.out_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="348" name="__module.layers.4.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="265" name="__module.layers.4.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5817,37 +4163,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="349" name="Constant_5552_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="30672996" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="350" name="Constant_5552" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="266" name="Constant_5934" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="61345892" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="351" name="__module.layers.4.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="267" name="__module.layers.4.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5862,14 +4188,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="444,input.47">
+				<port id="2" precision="FP32" names="409,input.27">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="352" name="__module.layers.4/aten::add/Add" type="Add" version="opset1">
+		<layer id="268" name="__module.layers.4/aten::add/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5884,22 +4210,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="446,residual.19">
+				<port id="2" precision="FP32" names="411,residual.19">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="353" name="__module.layers.4.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="269" name="__module.layers.4.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="354" name="__module.layers.4.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="270" name="__module.layers.4.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5919,37 +4245,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="355" name="Constant_5553_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="30674020" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="356" name="Constant_5553" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="271" name="Constant_5935" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="61347940" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="357" name="__module.layers.4.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="272" name="__module.layers.4.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -5971,37 +4277,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="358" name="Constant_5554_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="30675044" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="359" name="Constant_5554" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="273" name="Constant_5936" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="61349988" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="360" name="__module.layers.4.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="274" name="__module.layers.4.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6016,41 +4302,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="450">
+				<port id="2" precision="FP32" names="415">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="361" name="self.layers.4.fc1.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="2048, 512" offset="30676068" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.4.fc1.weight">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="362" name="self.layers.4.fc1.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="275" name="self.layers.4.fc1.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="2048, 512" offset="61352036" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.4.fc1.weight">
 					<dim>2048</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="363" name="__module.layers.4.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="276" name="__module.layers.4.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6071,37 +4339,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="364" name="Constant_5555_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 2048" offset="32773220" size="4096" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="365" name="Constant_5555" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="277" name="Constant_5937" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 2048" offset="65546340" size="8192" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="366" name="__module.layers.4.fc1/aten::linear/Add" type="Add" version="opset1">
+		<layer id="278" name="__module.layers.4.fc1/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6116,14 +4364,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="453">
+				<port id="2" precision="FP32" names="418">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="367" name="__module.layers.4.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
+		<layer id="279" name="__module.layers.4.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6133,41 +4381,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="1" precision="FP32" names="454,input.49">
+				<port id="1" precision="FP32" names="419,input.29">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="368" name="self.layers.4.fc2.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 2048" offset="32777316" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.4.fc2.weight">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="369" name="self.layers.4.fc2.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="280" name="self.layers.4.fc2.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 2048" offset="65554532" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.4.fc2.weight">
 					<dim>512</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="370" name="__module.layers.4.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="281" name="__module.layers.4.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6188,37 +4418,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="371" name="Constant_5556_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="34874468" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="372" name="Constant_5556" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="282" name="Constant_5938" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="69748836" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="373" name="__module.layers.4.fc2/aten::linear/Add" type="Add" version="opset1">
+		<layer id="283" name="__module.layers.4.fc2/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6233,14 +4443,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="458,input.51">
+				<port id="2" precision="FP32" names="423,input.31">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="374" name="__module.layers.4/aten::add/Add_1" type="Add" version="opset1">
+		<layer id="284" name="__module.layers.4/aten::add/Add_1" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6255,22 +4465,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="460,residual.21">
+				<port id="2" precision="FP32" names="425,residual.21">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="375" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="285" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="376" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="286" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6290,37 +4500,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="377" name="Constant_5557_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="34875492" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="378" name="Constant_5557" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="287" name="Constant_5939" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="69750884" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="379" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="288" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6342,37 +4532,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="380" name="Constant_5558_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="34876516" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="381" name="Constant_5558" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="289" name="Constant_5940" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="69752932" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="382" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="290" name="__module.layers.5.self_attn_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6387,41 +4557,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="485,hidden_states.31">
+				<port id="2" precision="FP32" names="448,hidden_states.31">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="383" name="Constant_5447_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="34877540" size="524288" />
+		<layer id="291" name="self.layers.5.self_attn.q_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="69754980" size="1048576" />
 			<output>
-				<port id="0" precision="FP16">
+				<port id="0" precision="FP32" names="self.layers.5.self_attn.q_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="384" name="Constant_5447" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="385" name="Multiply_5427" type="MatMul" version="opset1">
+		<layer id="292" name="__module.layers.5.self_attn.q_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6442,37 +4594,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="386" name="Constant_5559_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="35401828" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="387" name="Constant_5559" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="293" name="Constant_5941" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="70803556" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="388" name="__module.layers.5.self_attn/aten::mul/Multiply" type="Add" version="opset1">
+		<layer id="294" name="__module.layers.5.self_attn.q_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6487,25 +4619,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="495,tensor.31">
+				<port id="2" precision="FP32" names="457,tensor.31">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="389" name="Constant_5694" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="295" name="Constant_6076" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="390" name="__module.layers.5.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
+		<layer id="296" name="__module.layers.5.self_attn/aten::view/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6518,7 +4647,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="497">
+				<port id="2" precision="FP32" names="459">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -6526,15 +4655,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="391" name="__module.layers.5.self_attn/aten::transpose/Constant" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="297" name="__module.layers.5.self_attn/aten::transpose/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="392" name="__module.layers.5.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
+		<layer id="298" name="__module.layers.5.self_attn/aten::transpose/Transpose" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -6547,7 +4676,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="498">
+				<port id="2" precision="FP32" names="460">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -6555,34 +4684,16 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="393" name="self.layers.5.self_attn.k_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="35402852" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.5.self_attn.k_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="394" name="self.layers.5.self_attn.k_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="299" name="self.layers.5.self_attn.k_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="70805604" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.5.self_attn.k_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="395" name="__module.layers.5.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="300" name="__module.layers.5.self_attn.k_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6596,25 +4707,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="501,tensor.33">
+				<port id="2" precision="FP32" names="463,tensor.33">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="396" name="Constant_5695" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="301" name="Constant_6077" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="397" name="__module.layers.5.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
+		<layer id="302" name="__module.layers.5.self_attn/aten::view/Reshape_1" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6627,7 +4735,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="503">
+				<port id="2" precision="FP32" names="465">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -6635,15 +4743,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="398" name="__module.layers.5.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="303" name="__module.layers.5.self_attn/aten::transpose/Constant_1" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="399" name="__module.layers.5.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
+		<layer id="304" name="__module.layers.5.self_attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -6656,86 +4764,24 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="504">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="400" name="__module.layers.5.self_attn/aten::matmul/MatMul" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="true" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>64</dim>
-				</port>
-				<port id="1" precision="FP32">
+				<port id="2" precision="FP32" names="466">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
-			</input>
-			<output>
-				<port id="2" precision="FP32" names="514,input.53">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="401" name="__module.layers.5.self_attn/aten::softmax/Softmax" type="SoftMax" version="opset8">
-			<data axis="-1" />
-			<input>
-				<port id="0" precision="FP32">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</input>
-			<output>
-				<port id="1" precision="FP32" names="515,input.55">
-					<dim>-1</dim>
-					<dim>8</dim>
-					<dim>1500</dim>
-					<dim>1500</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="402" name="self.layers.5.self_attn.v_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="35927140" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.5.self_attn.v_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
 			</output>
 		</layer>
-		<layer id="403" name="self.layers.5.self_attn.v_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="305" name="self.layers.5.self_attn.v_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="71854180" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.5.self_attn.v_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="404" name="__module.layers.5.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="306" name="__module.layers.5.self_attn.v_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6756,37 +4802,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="405" name="Constant_5560_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="36451428" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="406" name="Constant_5560" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="307" name="Constant_5942" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="72902756" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="407" name="__module.layers.5.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="308" name="__module.layers.5.self_attn.v_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6801,25 +4827,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="508,tensor">
+				<port id="2" precision="FP32" names="470,tensor">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="408" name="Constant_5696" type="Const" version="opset1">
-			<data element_type="i64" shape="4" offset="3884060" size="32" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="309" name="Constant_6078" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="7768092" size="32" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="409" name="__module.layers.5.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
+		<layer id="310" name="__module.layers.5.self_attn/aten::view/Reshape_2" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6832,7 +4855,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="510">
+				<port id="2" precision="FP32" names="472">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -6840,15 +4863,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="410" name="__module.layers.5.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="311" name="__module.layers.5.self_attn/aten::transpose/Constant_2" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="411" name="__module.layers.5.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
+		<layer id="312" name="__module.layers.5.self_attn/aten::transpose/Transpose_2" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -6861,7 +4884,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="511">
+				<port id="2" precision="FP32" names="473">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -6869,14 +4892,14 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="412" name="__module.layers.5.self_attn/aten::matmul/MatMul_1" type="MatMul" version="opset1">
-			<data transpose_a="false" transpose_b="false" />
+		<layer id="313" name="__module.layers.5.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>8</dim>
 					<dim>1500</dim>
-					<dim>1500</dim>
+					<dim>64</dim>
 				</port>
 				<port id="1" precision="FP32">
 					<dim>-1</dim>
@@ -6884,9 +4907,15 @@
 					<dim>1500</dim>
 					<dim>64</dim>
 				</port>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>8</dim>
+					<dim>1500</dim>
+					<dim>64</dim>
+				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="517,attn_output.21">
+				<port id="3" precision="FP32" names="475,attn_output.21">
 					<dim>-1</dim>
 					<dim>8</dim>
 					<dim>1500</dim>
@@ -6894,15 +4923,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="413" name="__module.layers.5.self_attn/aten::transpose/Constant_4" type="Const" version="opset1">
-			<data element_type="i32" shape="4" offset="3884092" size="16" />
+		<layer id="314" name="__module.layers.5.self_attn/aten::transpose/Constant_3" type="Const" version="opset1">
+			<data element_type="i32" shape="4" offset="7768124" size="16" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>4</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="414" name="__module.layers.5.self_attn/aten::transpose/Transpose_4" type="Transpose" version="opset1">
+		<layer id="315" name="__module.layers.5.self_attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -6915,7 +4944,7 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="518,attn_output">
+				<port id="2" precision="FP32" names="476,attn_output">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>8</dim>
@@ -6923,18 +4952,15 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="415" name="Constant_5697" type="Const" version="opset1">
-			<data element_type="i64" shape="3" offset="4933708" size="24" />
-			<rt_info>
-				<attribute name="precise" version="0" />
-			</rt_info>
+		<layer id="316" name="Constant_6079" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="9867340" size="24" />
 			<output>
 				<port id="0" precision="I64">
 					<dim>3</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="416" name="__module.layers.5.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
+		<layer id="317" name="__module.layers.5.self_attn/aten::reshape/Reshape" type="Reshape" version="opset1">
 			<data special_zero="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -6948,41 +4974,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="520">
+				<port id="2" precision="FP32" names="478">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="417" name="self.layers.5.self_attn.out_proj.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 512" offset="36452452" size="524288" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.5.self_attn.out_proj.weight">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="418" name="self.layers.5.self_attn.out_proj.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="318" name="self.layers.5.self_attn.out_proj.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 512" offset="72904804" size="1048576" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.5.self_attn.out_proj.weight">
 					<dim>512</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="419" name="__module.layers.5.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="319" name="__module.layers.5.self_attn.out_proj/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7003,37 +5011,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="420" name="Constant_5561_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="36976740" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="421" name="Constant_5561" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="320" name="Constant_5943" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="73953380" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="422" name="__module.layers.5.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
+		<layer id="321" name="__module.layers.5.self_attn.out_proj/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7048,14 +5036,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="523,input.57">
+				<port id="2" precision="FP32" names="481,input.33">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="423" name="__module.layers.5/aten::add/Add" type="Add" version="opset1">
+		<layer id="322" name="__module.layers.5/aten::add/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7070,22 +5058,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="525,residual">
+				<port id="2" precision="FP32" names="483,residual">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="424" name="__module.layers.5.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="323" name="__module.layers.5.final_layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="425" name="__module.layers.5.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="324" name="__module.layers.5.final_layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7105,37 +5093,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="426" name="Constant_5562_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="36977764" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="427" name="Constant_5562" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="325" name="Constant_5944" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="73955428" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="428" name="__module.layers.5.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="326" name="__module.layers.5.final_layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7157,37 +5125,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="429" name="Constant_5563_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="36978788" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="430" name="Constant_5563" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="327" name="Constant_5945" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="73957476" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="431" name="__module.layers.5.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="328" name="__module.layers.5.final_layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7202,41 +5150,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="529">
+				<port id="2" precision="FP32" names="487">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="432" name="self.layers.5.fc1.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="2048, 512" offset="36979812" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.5.fc1.weight">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="433" name="self.layers.5.fc1.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>2048</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="329" name="self.layers.5.fc1.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="2048, 512" offset="73959524" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.5.fc1.weight">
 					<dim>2048</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="434" name="__module.layers.5.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="330" name="__module.layers.5.fc1/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7257,37 +5187,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="435" name="Constant_5564_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 2048" offset="39076964" size="4096" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="436" name="Constant_5564" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="331" name="Constant_5946" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 2048" offset="78153828" size="8192" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="437" name="__module.layers.5.fc1/aten::linear/Add" type="Add" version="opset1">
+		<layer id="332" name="__module.layers.5.fc1/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7302,14 +5212,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="532">
+				<port id="2" precision="FP32" names="490">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="438" name="__module.layers.5.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
+		<layer id="333" name="__module.layers.5.activation_fn/aten::gelu/Gelu" type="Gelu" version="opset7">
 			<data approximation_mode="ERF" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7319,41 +5229,23 @@
 				</port>
 			</input>
 			<output>
-				<port id="1" precision="FP32" names="533,input.59">
+				<port id="1" precision="FP32" names="491,input.35">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="439" name="self.layers.5.fc2.weight_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="512, 2048" offset="39081060" size="2097152" />
-			<output>
-				<port id="0" precision="FP16" names="self.layers.5.fc2.weight">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="440" name="self.layers.5.fc2.weight" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>512</dim>
-					<dim>2048</dim>
-				</port>
-			</input>
+		<layer id="334" name="self.layers.5.fc2.weight" type="Const" version="opset1">
+			<data element_type="f32" shape="512, 2048" offset="78162020" size="4194304" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32" names="self.layers.5.fc2.weight">
 					<dim>512</dim>
 					<dim>2048</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="441" name="__module.layers.5.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
+		<layer id="335" name="__module.layers.5.fc2/aten::linear/MatMul" type="MatMul" version="opset1">
 			<data transpose_a="false" transpose_b="true" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7374,37 +5266,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="442" name="Constant_5565_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="41178212" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="443" name="Constant_5565" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="336" name="Constant_5947" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="82356324" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="444" name="__module.layers.5.fc2/aten::linear/Add" type="Add" version="opset1">
+		<layer id="337" name="__module.layers.5.fc2/aten::linear/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7419,14 +5291,14 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="537,input">
+				<port id="2" precision="FP32" names="495,input">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="445" name="__module.layers.5/aten::add/Add_1" type="Add" version="opset1">
+		<layer id="338" name="__module.layers.5/aten::add/Add_1" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7441,22 +5313,22 @@
 				</port>
 			</input>
 			<output>
-				<port id="2" precision="FP32" names="539">
+				<port id="2" precision="FP32" names="497">
 					<dim>-1</dim>
 					<dim>1500</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="446" name="__module.layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
-			<data element_type="i32" shape="1" offset="3356696" size="4" />
+		<layer id="339" name="__module.layer_norm/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="6713368" size="4" />
 			<output>
 				<port id="0" precision="I32">
 					<dim>1</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="447" name="__module.layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
+		<layer id="340" name="__module.layer_norm/aten::layer_norm/MVN" type="MVN" version="opset6">
 			<data eps="9.9999997473787516e-06" normalize_variance="true" eps_mode="INSIDE_SQRT" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7476,37 +5348,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="448" name="Constant_5566_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="41179236" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="449" name="Constant_5566" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="341" name="Constant_5948" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="82358372" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="450" name="__module.layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+		<layer id="342" name="__module.layer_norm/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7528,37 +5380,17 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="451" name="Constant_5567_compressed" type="Const" version="opset1">
-			<data element_type="f16" shape="1, 1, 512" offset="41180260" size="1024" />
-			<output>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</output>
-		</layer>
-		<layer id="452" name="Constant_5567" type="Convert" version="opset1">
-			<data destination_type="f32" />
-			<rt_info>
-				<attribute name="decompression" version="0" />
-			</rt_info>
-			<input>
-				<port id="0" precision="FP16">
-					<dim>1</dim>
-					<dim>1</dim>
-					<dim>512</dim>
-				</port>
-			</input>
+		<layer id="343" name="Constant_5949" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 512" offset="82360420" size="2048" />
 			<output>
-				<port id="1" precision="FP32">
+				<port id="0" precision="FP32">
 					<dim>1</dim>
 					<dim>1</dim>
 					<dim>512</dim>
 				</port>
 			</output>
 		</layer>
-		<layer id="453" name="__module.layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
+		<layer id="344" name="__module.layer_norm/aten::layer_norm/Add" type="Add" version="opset1">
 			<data auto_broadcast="numpy" />
 			<input>
 				<port id="0" precision="FP32">
@@ -7580,7 +5412,7 @@
 				</port>
 			</output>
 		</layer>
-		<layer id="454" name="Result_3137" type="Result" version="opset1">
+		<layer id="345" name="Result_1938" type="Result" version="opset1" output_names="last_hidden_state">
 			<input>
 				<port id="0" precision="FP32">
 					<dim>-1</dim>
@@ -7591,496 +5423,387 @@
 		</layer>
 	</layers>
 	<edges>
-		<edge from-layer="0" from-port="0" to-layer="3" to-port="0" />
-		<edge from-layer="1" from-port="0" to-layer="2" to-port="0" />
-		<edge from-layer="2" from-port="1" to-layer="3" to-port="1" />
-		<edge from-layer="3" from-port="2" to-layer="6" to-port="0" />
-		<edge from-layer="4" from-port="0" to-layer="5" to-port="0" />
-		<edge from-layer="5" from-port="1" to-layer="6" to-port="1" />
-		<edge from-layer="6" from-port="2" to-layer="7" to-port="0" />
-		<edge from-layer="7" from-port="1" to-layer="10" to-port="0" />
-		<edge from-layer="8" from-port="0" to-layer="9" to-port="0" />
-		<edge from-layer="9" from-port="1" to-layer="10" to-port="1" />
-		<edge from-layer="10" from-port="2" to-layer="13" to-port="0" />
-		<edge from-layer="11" from-port="0" to-layer="12" to-port="0" />
-		<edge from-layer="12" from-port="1" 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="0" from-port="0" to-layer="2" to-port="0" />
+		<edge from-layer="1" from-port="0" to-layer="2" to-port="1" />
+		<edge from-layer="2" from-port="2" to-layer="4" to-port="0" />
+		<edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
+		<edge from-layer="4" from-port="2" to-layer="5" to-port="0" />
+		<edge from-layer="5" from-port="1" to-layer="7" to-port="0" />
+		<edge from-layer="6" from-port="0" to-layer="7" to-port="1" />
+		<edge from-layer="7" from-port="2" to-layer="9" to-port="0" />
+		<edge from-layer="8" from-port="0" to-layer="9" to-port="1" />
+		<edge from-layer="9" from-port="2" to-layer="10" to-port="0" />
+		<edge from-layer="10" from-port="1" to-layer="12" to-port="0" />
+		<edge from-layer="11" from-port="0" to-layer="12" to-port="1" />
+		<edge from-layer="12" from-port="2" to-layer="14" to-port="0" />
+		<edge from-layer="13" from-port="0" to-layer="14" to-port="1" />
+		<edge from-layer="14" from-port="2" to-layer="52" to-port="0" />
+		<edge from-layer="14" from-port="2" to-layer="16" to-port="0" />
 		<edge from-layer="15" from-port="0" to-layer="16" to-port="1" />
-		<edge from-layer="16" from-port="2" to-layer="19" to-port="0" />
-		<edge from-layer="17" from-port="0" to-layer="18" to-port="0" />
-		<edge from-layer="18" from-port="1" to-layer="19" to-port="1" />
-		<edge from-layer="19" from-port="2" to-layer="68" to-port="0" />
-		<edge from-layer="19" from-port="2" to-layer="21" to-port="0" />
-		<edge from-layer="20" from-port="0" to-layer="21" to-port="1" />
-		<edge from-layer="21" from-port="2" to-layer="24" to-port="0" />
-		<edge from-layer="22" from-port="0" to-layer="23" to-port="0" />
-		<edge from-layer="23" from-port="1" to-layer="24" to-port="1" />
-		<edge from-layer="24" from-port="2" to-layer="27" to-port="0" />
-		<edge from-layer="25" from-port="0" to-layer="26" to-port="0" />
-		<edge from-layer="26" from-port="1" to-layer="27" to-port="1" />
-		<edge from-layer="27" from-port="2" to-layer="49" to-port="0" />
-		<edge from-layer="27" from-port="2" to-layer="40" to-port="0" />
-		<edge from-layer="27" from-port="2" to-layer="30" to-port="0" />
-		<edge from-layer="28" from-port="0" to-layer="29" to-port="0" />
-		<edge from-layer="29" from-port="1" to-layer="30" to-port="1" />
-		<edge from-layer="30" from-port="2" to-layer="33" to-port="0" />
-		<edge from-layer="31" from-port="0" to-layer="32" to-port="0" />
-		<edge from-layer="32" from-port="1" to-layer="33" to-port="1" />
-		<edge from-layer="33" from-port="2" to-layer="35" to-port="0" />
-		<edge from-layer="34" from-port="0" 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="45" 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="16" from-port="2" to-layer="18" 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="20" from-port="2" to-layer="36" to-port="0" />
+		<edge from-layer="20" from-port="2" to-layer="30" 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="26" to-port="0" />
+		<edge from-layer="25" from-port="0" to-layer="26" to-port="1" />
+		<edge from-layer="26" from-port="2" to-layer="28" 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="43" 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="34" to-port="0" />
+		<edge from-layer="33" from-port="0" to-layer="34" to-port="1" />
+		<edge from-layer="34" from-port="2" to-layer="43" to-port="1" />
+		<edge from-layer="35" from-port="0" to-layer="36" to-port="1" />
+		<edge from-layer="36" from-port="2" to-layer="38" to-port="0" />
+		<edge from-layer="37" from-port="0" to-layer="38" to-port="1" />
+		<edge from-layer="38" from-port="2" to-layer="40" to-port="0" />
+		<edge from-layer="39" from-port="0" 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="45" to-port="1" />
-		<edge from-layer="45" from-port="2" to-layer="46" to-port="0" />
-		<edge from-layer="46" from-port="1" to-layer="57" to-port="0" />
-		<edge from-layer="47" from-port="0" to-layer="48" to-port="0" />
-		<edge from-layer="48" from-port="1" to-layer="49" to-port="1" />
-		<edge from-layer="49" from-port="2" to-layer="52" to-port="0" />
-		<edge from-layer="50" from-port="0" to-layer="51" to-port="0" />
-		<edge from-layer="51" from-port="1" to-layer="52" to-port="1" />
+		<edge from-layer="42" from-port="2" to-layer="43" to-port="2" />
+		<edge from-layer="43" from-port="3" to-layer="45" to-port="0" />
+		<edge from-layer="44" from-port="0" to-layer="45" to-port="1" />
+		<edge from-layer="45" from-port="2" to-layer="47" to-port="0" />
+		<edge from-layer="46" from-port="0" to-layer="47" to-port="1" />
+		<edge from-layer="47" from-port="2" to-layer="49" to-port="0" />
+		<edge from-layer="48" from-port="0" to-layer="49" to-port="1" />
+		<edge from-layer="49" from-port="2" to-layer="51" to-port="0" />
+		<edge from-layer="50" from-port="0" to-layer="51" to-port="1" />
+		<edge from-layer="51" from-port="2" to-layer="52" to-port="1" />
 		<edge from-layer="52" from-port="2" to-layer="54" to-port="0" />
+		<edge from-layer="52" from-port="2" to-layer="68" to-port="0" />
 		<edge from-layer="53" from-port="0" to-layer="54" to-port="1" />
 		<edge from-layer="54" from-port="2" to-layer="56" to-port="0" />
 		<edge from-layer="55" from-port="0" to-layer="56" to-port="1" />
-		<edge from-layer="56" from-port="2" to-layer="57" to-port="1" />
-		<edge from-layer="57" from-port="2" to-layer="59" to-port="0" />
-		<edge from-layer="58" from-port="0" to-layer="59" to-port="1" />
-		<edge from-layer="59" from-port="2" to-layer="61" to-port="0" />
-		<edge from-layer="60" from-port="0" to-layer="61" to-port="1" />
-		<edge from-layer="61" from-port="2" to-layer="64" to-port="0" />
-		<edge from-layer="62" from-port="0" to-layer="63" to-port="0" />
-		<edge from-layer="63" from-port="1" to-layer="64" to-port="1" />
-		<edge from-layer="64" from-port="2" to-layer="67" to-port="0" />
-		<edge from-layer="65" from-port="0" to-layer="66" to-port="0" />
-		<edge from-layer="66" from-port="1" to-layer="67" to-port="1" />
+		<edge from-layer="56" from-port="2" to-layer="58" to-port="0" />
+		<edge from-layer="57" from-port="0" to-layer="58" to-port="1" />
+		<edge from-layer="58" from-port="2" to-layer="60" to-port="0" />
+		<edge from-layer="59" from-port="0" 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="63" from-port="1" to-layer="65" to-port="0" />
+		<edge from-layer="64" from-port="0" to-layer="65" to-port="1" />
+		<edge from-layer="65" from-port="2" to-layer="67" to-port="0" />
+		<edge from-layer="66" from-port="0" to-layer="67" to-port="1" />
 		<edge from-layer="67" from-port="2" to-layer="68" to-port="1" />
-		<edge from-layer="68" from-port="2" to-layer="90" to-port="0" />
 		<edge from-layer="68" from-port="2" to-layer="70" to-port="0" />
+		<edge from-layer="68" from-port="2" to-layer="106" to-port="0" />
 		<edge from-layer="69" from-port="0" to-layer="70" to-port="1" />
-		<edge from-layer="70" from-port="2" to-layer="73" to-port="0" />
-		<edge from-layer="71" from-port="0" to-layer="72" to-port="0" />
-		<edge from-layer="72" from-port="1" to-layer="73" to-port="1" />
-		<edge from-layer="73" from-port="2" to-layer="76" to-port="0" />
-		<edge from-layer="74" from-port="0" to-layer="75" to-port="0" />
-		<edge from-layer="75" from-port="1" to-layer="76" to-port="1" />
-		<edge from-layer="76" from-port="2" to-layer="79" to-port="0" />
-		<edge from-layer="77" from-port="0" to-layer="78" to-port="0" />
-		<edge from-layer="78" from-port="1" to-layer="79" to-port="1" />
-		<edge from-layer="79" from-port="2" to-layer="82" to-port="0" />
-		<edge from-layer="80" from-port="0" to-layer="81" to-port="0" />
-		<edge from-layer="81" from-port="1" to-layer="82" to-port="1" />
-		<edge from-layer="82" from-port="2" to-layer="83" to-port="0" />
-		<edge from-layer="83" from-port="1" to-layer="86" to-port="0" />
-		<edge from-layer="84" from-port="0" to-layer="85" to-port="0" />
-		<edge from-layer="85" from-port="1" to-layer="86" to-port="1" />
-		<edge from-layer="86" from-port="2" to-layer="89" to-port="0" />
-		<edge from-layer="87" from-port="0" to-layer="88" to-port="0" />
-		<edge from-layer="88" from-port="1" to-layer="89" to-port="1" />
-		<edge from-layer="89" from-port="2" to-layer="90" to-port="1" />
+		<edge from-layer="70" from-port="2" to-layer="72" to-port="0" />
+		<edge from-layer="71" from-port="0" to-layer="72" to-port="1" />
+		<edge from-layer="72" from-port="2" to-layer="74" to-port="0" />
+		<edge from-layer="73" from-port="0" to-layer="74" to-port="1" />
+		<edge from-layer="74" from-port="2" to-layer="84" to-port="0" />
+		<edge from-layer="74" from-port="2" to-layer="90" to-port="0" />
+		<edge from-layer="74" from-port="2" to-layer="76" to-port="0" />
+		<edge from-layer="75" from-port="0" to-layer="76" to-port="1" />
+		<edge from-layer="76" from-port="2" to-layer="78" to-port="0" />
+		<edge from-layer="77" from-port="0" to-layer="78" to-port="1" />
+		<edge from-layer="78" from-port="2" to-layer="80" to-port="0" />
+		<edge from-layer="79" from-port="0" to-layer="80" to-port="1" />
+		<edge from-layer="80" from-port="2" to-layer="82" to-port="0" />
+		<edge from-layer="81" from-port="0" to-layer="82" to-port="1" />
+		<edge from-layer="82" from-port="2" to-layer="97" to-port="0" />
+		<edge from-layer="83" from-port="0" to-layer="84" to-port="1" />
+		<edge from-layer="84" from-port="2" to-layer="86" to-port="0" />
+		<edge from-layer="85" from-port="0" to-layer="86" to-port="1" />
+		<edge from-layer="86" from-port="2" to-layer="88" to-port="0" />
+		<edge from-layer="87" from-port="0" to-layer="88" to-port="1" />
+		<edge from-layer="88" from-port="2" to-layer="97" to-port="1" />
+		<edge from-layer="89" from-port="0" to-layer="90" to-port="1" />
 		<edge from-layer="90" from-port="2" to-layer="92" to-port="0" />
-		<edge from-layer="90" from-port="2" to-layer="139" to-port="0" />
 		<edge from-layer="91" from-port="0" to-layer="92" to-port="1" />
-		<edge from-layer="92" from-port="2" to-layer="95" to-port="0" />
-		<edge from-layer="93" from-port="0" to-layer="94" to-port="0" />
-		<edge from-layer="94" from-port="1" to-layer="95" to-port="1" />
-		<edge from-layer="95" from-port="2" to-layer="98" to-port="0" />
-		<edge from-layer="96" from-port="0" to-layer="97" to-port="0" />
-		<edge from-layer="97" from-port="1" to-layer="98" to-port="1" />
-		<edge from-layer="98" from-port="2" to-layer="120" to-port="0" />
-		<edge from-layer="98" from-port="2" to-layer="111" to-port="0" />
-		<edge from-layer="98" from-port="2" to-layer="101" to-port="0" />
-		<edge from-layer="99" from-port="0" to-layer="100" to-port="0" />
-		<edge from-layer="100" from-port="1" to-layer="101" to-port="1" />
-		<edge from-layer="101" from-port="2" to-layer="104" to-port="0" />
-		<edge from-layer="102" from-port="0" to-layer="103" to-port="0" />
-		<edge from-layer="103" from-port="1" 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="92" from-port="2" to-layer="94" to-port="0" />
+		<edge from-layer="93" from-port="0" to-layer="94" to-port="1" />
+		<edge from-layer="94" from-port="2" to-layer="96" to-port="0" />
+		<edge from-layer="95" from-port="0" to-layer="96" to-port="1" />
+		<edge from-layer="96" from-port="2" to-layer="97" to-port="2" />
+		<edge from-layer="97" from-port="3" to-layer="99" to-port="0" />
+		<edge from-layer="98" from-port="0" to-layer="99" to-port="1" />
+		<edge from-layer="99" from-port="2" to-layer="101" to-port="0" />
+		<edge from-layer="100" from-port="0" to-layer="101" to-port="1" />
+		<edge from-layer="101" from-port="2" to-layer="103" to-port="0" />
+		<edge from-layer="102" from-port="0" to-layer="103" to-port="1" />
+		<edge from-layer="103" from-port="2" to-layer="105" to-port="0" />
+		<edge from-layer="104" from-port="0" to-layer="105" to-port="1" />
+		<edge from-layer="105" from-port="2" to-layer="106" to-port="1" />
 		<edge from-layer="106" from-port="2" to-layer="108" to-port="0" />
+		<edge from-layer="106" from-port="2" to-layer="122" 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="116" to-port="0" />
-		<edge from-layer="109" from-port="0" to-layer="110" to-port="0" />
-		<edge from-layer="110" from-port="1" to-layer="111" to-port="1" />
-		<edge from-layer="111" from-port="2" to-layer="113" to-port="0" />
-		<edge from-layer="112" from-port="0" to-layer="113" to-port="1" />
-		<edge from-layer="113" from-port="2" to-layer="115" to-port="0" />
-		<edge from-layer="114" from-port="0" to-layer="115" to-port="1" />
-		<edge from-layer="115" from-port="2" to-layer="116" to-port="1" />
+		<edge from-layer="108" from-port="2" to-layer="110" to-port="0" />
+		<edge from-layer="109" from-port="0" to-layer="110" to-port="1" />
+		<edge from-layer="110" from-port="2" to-layer="112" to-port="0" />
+		<edge from-layer="111" from-port="0" to-layer="112" to-port="1" />
+		<edge from-layer="112" from-port="2" to-layer="114" to-port="0" />
+		<edge from-layer="113" from-port="0" to-layer="114" to-port="1" />
+		<edge from-layer="114" from-port="2" to-layer="116" to-port="0" />
+		<edge from-layer="115" from-port="0" to-layer="116" to-port="1" />
 		<edge from-layer="116" from-port="2" to-layer="117" to-port="0" />
-		<edge from-layer="117" from-port="1" to-layer="128" to-port="0" />
-		<edge from-layer="118" from-port="0" to-layer="119" to-port="0" />
-		<edge from-layer="119" from-port="1" to-layer="120" to-port="1" />
-		<edge from-layer="120" from-port="2" to-layer="123" to-port="0" />
-		<edge from-layer="121" from-port="0" to-layer="122" to-port="0" />
-		<edge from-layer="122" from-port="1" to-layer="123" to-port="1" />
-		<edge from-layer="123" from-port="2" to-layer="125" to-port="0" />
-		<edge from-layer="124" from-port="0" to-layer="125" to-port="1" />
-		<edge from-layer="125" from-port="2" to-layer="127" to-port="0" />
-		<edge from-layer="126" from-port="0" to-layer="127" to-port="1" />
-		<edge from-layer="127" from-port="2" to-layer="128" to-port="1" />
+		<edge from-layer="117" from-port="1" to-layer="119" to-port="0" />
+		<edge from-layer="118" from-port="0" to-layer="119" to-port="1" />
+		<edge from-layer="119" from-port="2" to-layer="121" to-port="0" />
+		<edge from-layer="120" from-port="0" to-layer="121" to-port="1" />
+		<edge from-layer="121" from-port="2" to-layer="122" to-port="1" />
+		<edge from-layer="122" from-port="2" to-layer="124" to-port="0" />
+		<edge from-layer="122" from-port="2" to-layer="160" to-port="0" />
+		<edge from-layer="123" from-port="0" to-layer="124" to-port="1" />
+		<edge from-layer="124" from-port="2" to-layer="126" to-port="0" />
+		<edge from-layer="125" from-port="0" to-layer="126" to-port="1" />
+		<edge from-layer="126" from-port="2" to-layer="128" to-port="0" />
+		<edge from-layer="127" from-port="0" to-layer="128" to-port="1" />
 		<edge from-layer="128" from-port="2" to-layer="130" to-port="0" />
+		<edge from-layer="128" from-port="2" to-layer="144" to-port="0" />
+		<edge from-layer="128" from-port="2" to-layer="138" to-port="0" />
 		<edge from-layer="129" from-port="0" to-layer="130" to-port="1" />
 		<edge from-layer="130" from-port="2" to-layer="132" to-port="0" />
 		<edge from-layer="131" from-port="0" to-layer="132" to-port="1" />
-		<edge from-layer="132" from-port="2" to-layer="135" to-port="0" />
-		<edge from-layer="133" from-port="0" to-layer="134" to-port="0" />
-		<edge from-layer="134" from-port="1" to-layer="135" to-port="1" />
-		<edge from-layer="135" from-port="2" to-layer="138" to-port="0" />
-		<edge from-layer="136" from-port="0" to-layer="137" to-port="0" />
-		<edge from-layer="137" from-port="1" to-layer="138" to-port="1" />
-		<edge from-layer="138" from-port="2" to-layer="139" to-port="1" />
-		<edge from-layer="139" from-port="2" to-layer="141" to-port="0" />
-		<edge from-layer="139" from-port="2" to-layer="161" to-port="0" />
-		<edge from-layer="140" from-port="0" to-layer="141" to-port="1" />
-		<edge from-layer="141" from-port="2" to-layer="144" to-port="0" />
-		<edge from-layer="142" from-port="0" to-layer="143" to-port="0" />
-		<edge from-layer="143" from-port="1" to-layer="144" to-port="1" />
-		<edge from-layer="144" from-port="2" to-layer="147" to-port="0" />
-		<edge from-layer="145" from-port="0" to-layer="146" to-port="0" />
-		<edge from-layer="146" from-port="1" to-layer="147" to-port="1" />
-		<edge from-layer="147" from-port="2" to-layer="150" to-port="0" />
-		<edge from-layer="148" from-port="0" to-layer="149" to-port="0" />
-		<edge from-layer="149" from-port="1" to-layer="150" to-port="1" />
-		<edge from-layer="150" from-port="2" to-layer="153" to-port="0" />
-		<edge from-layer="151" from-port="0" to-layer="152" to-port="0" />
-		<edge from-layer="152" from-port="1" to-layer="153" to-port="1" />
-		<edge from-layer="153" from-port="2" to-layer="154" to-port="0" />
-		<edge from-layer="154" from-port="1" to-layer="157" to-port="0" />
-		<edge from-layer="155" from-port="0" to-layer="156" to-port="0" />
-		<edge from-layer="156" from-port="1" to-layer="157" to-port="1" />
-		<edge from-layer="157" from-port="2" to-layer="160" to-port="0" />
-		<edge from-layer="158" from-port="0" to-layer="159" to-port="0" />
-		<edge from-layer="159" from-port="1" to-layer="160" to-port="1" />
-		<edge from-layer="160" from-port="2" to-layer="161" to-port="1" />
-		<edge from-layer="161" from-port="2" to-layer="210" to-port="0" />
-		<edge from-layer="161" from-port="2" to-layer="163" to-port="0" />
-		<edge from-layer="162" from-port="0" to-layer="163" to-port="1" />
-		<edge from-layer="163" from-port="2" to-layer="166" to-port="0" />
-		<edge from-layer="164" from-port="0" to-layer="165" to-port="0" />
-		<edge from-layer="165" from-port="1" to-layer="166" to-port="1" />
-		<edge from-layer="166" from-port="2" to-layer="169" to-port="0" />
-		<edge from-layer="167" from-port="0" to-layer="168" to-port="0" />
-		<edge from-layer="168" from-port="1" to-layer="169" to-port="1" />
-		<edge from-layer="169" from-port="2" to-layer="172" to-port="0" />
-		<edge from-layer="169" from-port="2" to-layer="191" to-port="0" />
-		<edge from-layer="169" from-port="2" to-layer="182" to-port="0" />
-		<edge from-layer="170" from-port="0" to-layer="171" to-port="0" />
-		<edge from-layer="171" from-port="1" to-layer="172" to-port="1" />
-		<edge from-layer="172" from-port="2" to-layer="175" to-port="0" />
-		<edge from-layer="173" from-port="0" to-layer="174" to-port="0" />
-		<edge from-layer="174" from-port="1" to-layer="175" to-port="1" />
-		<edge from-layer="175" from-port="2" to-layer="177" to-port="0" />
-		<edge from-layer="176" from-port="0" to-layer="177" to-port="1" />
-		<edge from-layer="177" from-port="2" to-layer="179" to-port="0" />
-		<edge from-layer="178" from-port="0" to-layer="179" to-port="1" />
-		<edge from-layer="179" from-port="2" to-layer="187" to-port="0" />
-		<edge from-layer="180" from-port="0" to-layer="181" to-port="0" />
-		<edge from-layer="181" from-port="1" to-layer="182" to-port="1" />
+		<edge from-layer="132" from-port="2" to-layer="134" to-port="0" />
+		<edge from-layer="133" from-port="0" to-layer="134" to-port="1" />
+		<edge from-layer="134" from-port="2" to-layer="136" to-port="0" />
+		<edge from-layer="135" from-port="0" to-layer="136" to-port="1" />
+		<edge from-layer="136" from-port="2" to-layer="151" to-port="0" />
+		<edge from-layer="137" from-port="0" to-layer="138" to-port="1" />
+		<edge from-layer="138" from-port="2" 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="142" to-port="0" />
+		<edge from-layer="141" from-port="0" to-layer="142" to-port="1" />
+		<edge from-layer="142" from-port="2" to-layer="151" to-port="1" />
+		<edge from-layer="143" from-port="0" to-layer="144" to-port="1" />
+		<edge from-layer="144" from-port="2" to-layer="146" to-port="0" />
+		<edge from-layer="145" from-port="0" to-layer="146" to-port="1" />
+		<edge from-layer="146" from-port="2" to-layer="148" to-port="0" />
+		<edge from-layer="147" from-port="0" to-layer="148" to-port="1" />
+		<edge from-layer="148" from-port="2" to-layer="150" to-port="0" />
+		<edge from-layer="149" from-port="0" to-layer="150" to-port="1" />
+		<edge from-layer="150" from-port="2" to-layer="151" to-port="2" />
+		<edge from-layer="151" from-port="3" to-layer="153" to-port="0" />
+		<edge from-layer="152" from-port="0" to-layer="153" to-port="1" />
+		<edge from-layer="153" from-port="2" to-layer="155" to-port="0" />
+		<edge from-layer="154" from-port="0" to-layer="155" to-port="1" />
+		<edge from-layer="155" from-port="2" to-layer="157" to-port="0" />
+		<edge from-layer="156" from-port="0" to-layer="157" to-port="1" />
+		<edge from-layer="157" from-port="2" to-layer="159" to-port="0" />
+		<edge from-layer="158" from-port="0" to-layer="159" to-port="1" />
+		<edge from-layer="159" from-port="2" to-layer="160" to-port="1" />
+		<edge from-layer="160" from-port="2" to-layer="176" to-port="0" />
+		<edge from-layer="160" from-port="2" to-layer="162" to-port="0" />
+		<edge from-layer="161" from-port="0" to-layer="162" to-port="1" />
+		<edge from-layer="162" from-port="2" to-layer="164" to-port="0" />
+		<edge from-layer="163" from-port="0" to-layer="164" to-port="1" />
+		<edge from-layer="164" from-port="2" to-layer="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="170" to-port="0" />
+		<edge from-layer="169" from-port="0" to-layer="170" to-port="1" />
+		<edge from-layer="170" from-port="2" to-layer="171" to-port="0" />
+		<edge from-layer="171" from-port="1" 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="175" to-port="0" />
+		<edge from-layer="174" from-port="0" to-layer="175" to-port="1" />
+		<edge from-layer="175" from-port="2" to-layer="176" to-port="1" />
+		<edge from-layer="176" from-port="2" to-layer="178" to-port="0" />
+		<edge from-layer="176" from-port="2" to-layer="214" to-port="0" />
+		<edge from-layer="177" from-port="0" to-layer="178" to-port="1" />
+		<edge from-layer="178" from-port="2" to-layer="180" to-port="0" />
+		<edge from-layer="179" from-port="0" to-layer="180" to-port="1" />
+		<edge from-layer="180" from-port="2" to-layer="182" to-port="0" />
+		<edge from-layer="181" from-port="0" to-layer="182" to-port="1" />
+		<edge from-layer="182" from-port="2" to-layer="198" to-port="0" />
+		<edge from-layer="182" from-port="2" to-layer="192" to-port="0" />
 		<edge from-layer="182" from-port="2" to-layer="184" to-port="0" />
 		<edge from-layer="183" from-port="0" to-layer="184" to-port="1" />
 		<edge from-layer="184" from-port="2" to-layer="186" 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="187" to-port="1" />
-		<edge from-layer="187" from-port="2" to-layer="188" to-port="0" />
-		<edge from-layer="188" from-port="1" to-layer="199" to-port="0" />
-		<edge from-layer="189" from-port="0" to-layer="190" to-port="0" />
-		<edge from-layer="190" from-port="1" to-layer="191" to-port="1" />
-		<edge from-layer="191" from-port="2" to-layer="194" to-port="0" />
-		<edge from-layer="192" from-port="0" to-layer="193" to-port="0" />
-		<edge from-layer="193" from-port="1" to-layer="194" to-port="1" />
+		<edge from-layer="186" from-port="2" to-layer="188" to-port="0" />
+		<edge from-layer="187" from-port="0" to-layer="188" to-port="1" />
+		<edge from-layer="188" from-port="2" to-layer="190" to-port="0" />
+		<edge from-layer="189" from-port="0" to-layer="190" to-port="1" />
+		<edge from-layer="190" from-port="2" to-layer="205" to-port="0" />
+		<edge from-layer="191" from-port="0" to-layer="192" to-port="1" />
+		<edge from-layer="192" from-port="2" to-layer="194" to-port="0" />
+		<edge from-layer="193" from-port="0" to-layer="194" to-port="1" />
 		<edge from-layer="194" from-port="2" to-layer="196" to-port="0" />
 		<edge from-layer="195" from-port="0" to-layer="196" to-port="1" />
-		<edge from-layer="196" from-port="2" to-layer="198" to-port="0" />
+		<edge from-layer="196" from-port="2" to-layer="205" 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="201" to-port="0" />
-		<edge from-layer="200" from-port="0" to-layer="201" to-port="1" />
-		<edge from-layer="201" from-port="2" to-layer="203" to-port="0" />
-		<edge from-layer="202" from-port="0" to-layer="203" to-port="1" />
-		<edge from-layer="203" from-port="2" to-layer="206" to-port="0" />
-		<edge from-layer="204" from-port="0" to-layer="205" to-port="0" />
-		<edge from-layer="205" from-port="1" to-layer="206" to-port="1" />
-		<edge from-layer="206" from-port="2" to-layer="209" to-port="0" />
-		<edge from-layer="207" from-port="0" to-layer="208" to-port="0" />
-		<edge from-layer="208" from-port="1" 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="232" to-port="0" />
-		<edge from-layer="210" from-port="2" to-layer="212" to-port="0" />
-		<edge from-layer="211" from-port="0" to-layer="212" to-port="1" />
-		<edge from-layer="212" from-port="2" to-layer="215" to-port="0" />
-		<edge from-layer="213" from-port="0" to-layer="214" to-port="0" />
-		<edge from-layer="214" from-port="1" to-layer="215" to-port="1" />
-		<edge from-layer="215" from-port="2" to-layer="218" to-port="0" />
-		<edge from-layer="216" from-port="0" to-layer="217" to-port="0" />
-		<edge from-layer="217" from-port="1" to-layer="218" to-port="1" />
-		<edge from-layer="218" from-port="2" to-layer="221" to-port="0" />
-		<edge from-layer="219" from-port="0" to-layer="220" to-port="0" />
-		<edge from-layer="220" from-port="1" to-layer="221" to-port="1" />
-		<edge from-layer="221" from-port="2" to-layer="224" to-port="0" />
-		<edge from-layer="222" from-port="0" to-layer="223" to-port="0" />
-		<edge from-layer="223" from-port="1" to-layer="224" to-port="1" />
+		<edge from-layer="198" from-port="2" to-layer="200" to-port="0" />
+		<edge from-layer="199" from-port="0" to-layer="200" to-port="1" />
+		<edge from-layer="200" from-port="2" to-layer="202" to-port="0" />
+		<edge from-layer="201" from-port="0" to-layer="202" to-port="1" />
+		<edge from-layer="202" from-port="2" to-layer="204" to-port="0" />
+		<edge from-layer="203" from-port="0" to-layer="204" to-port="1" />
+		<edge from-layer="204" from-port="2" to-layer="205" to-port="2" />
+		<edge from-layer="205" from-port="3" 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="211" to-port="0" />
+		<edge from-layer="210" from-port="0" to-layer="211" to-port="1" />
+		<edge from-layer="211" from-port="2" to-layer="213" to-port="0" />
+		<edge from-layer="212" from-port="0" to-layer="213" to-port="1" />
+		<edge from-layer="213" from-port="2" to-layer="214" to-port="1" />
+		<edge from-layer="214" from-port="2" to-layer="216" to-port="0" />
+		<edge from-layer="214" from-port="2" to-layer="230" to-port="0" />
+		<edge from-layer="215" from-port="0" to-layer="216" to-port="1" />
+		<edge from-layer="216" from-port="2" to-layer="218" to-port="0" />
+		<edge from-layer="217" from-port="0" to-layer="218" to-port="1" />
+		<edge from-layer="218" from-port="2" to-layer="220" to-port="0" />
+		<edge from-layer="219" from-port="0" to-layer="220" to-port="1" />
+		<edge from-layer="220" from-port="2" to-layer="222" to-port="0" />
+		<edge from-layer="221" from-port="0" to-layer="222" to-port="1" />
+		<edge from-layer="222" from-port="2" to-layer="224" to-port="0" />
+		<edge from-layer="223" from-port="0" to-layer="224" to-port="1" />
 		<edge from-layer="224" from-port="2" to-layer="225" to-port="0" />
-		<edge from-layer="225" from-port="1" to-layer="228" to-port="0" />
-		<edge from-layer="226" from-port="0" to-layer="227" to-port="0" />
-		<edge from-layer="227" from-port="1" to-layer="228" to-port="1" />
-		<edge from-layer="228" from-port="2" to-layer="231" to-port="0" />
-		<edge from-layer="229" from-port="0" to-layer="230" to-port="0" />
-		<edge from-layer="230" from-port="1" to-layer="231" to-port="1" />
-		<edge from-layer="231" from-port="2" to-layer="232" to-port="1" />
+		<edge from-layer="225" from-port="1" to-layer="227" to-port="0" />
+		<edge from-layer="226" from-port="0" to-layer="227" to-port="1" />
+		<edge from-layer="227" from-port="2" to-layer="229" to-port="0" />
+		<edge from-layer="228" from-port="0" to-layer="229" to-port="1" />
+		<edge from-layer="229" from-port="2" to-layer="230" to-port="1" />
+		<edge from-layer="230" from-port="2" to-layer="268" to-port="0" />
+		<edge from-layer="230" from-port="2" to-layer="232" to-port="0" />
+		<edge from-layer="231" from-port="0" to-layer="232" to-port="1" />
 		<edge from-layer="232" from-port="2" to-layer="234" to-port="0" />
-		<edge from-layer="232" from-port="2" to-layer="281" to-port="0" />
 		<edge from-layer="233" from-port="0" to-layer="234" to-port="1" />
-		<edge from-layer="234" from-port="2" to-layer="237" to-port="0" />
-		<edge from-layer="235" from-port="0" to-layer="236" to-port="0" />
-		<edge from-layer="236" from-port="1" to-layer="237" to-port="1" />
-		<edge from-layer="237" from-port="2" to-layer="240" to-port="0" />
-		<edge from-layer="238" from-port="0" to-layer="239" to-port="0" />
-		<edge from-layer="239" from-port="1" to-layer="240" to-port="1" />
-		<edge from-layer="240" from-port="2" to-layer="253" to-port="0" />
-		<edge from-layer="240" from-port="2" to-layer="262" to-port="0" />
-		<edge from-layer="240" from-port="2" to-layer="243" to-port="0" />
-		<edge from-layer="241" from-port="0" to-layer="242" to-port="0" />
-		<edge from-layer="242" from-port="1" to-layer="243" to-port="1" />
-		<edge from-layer="243" from-port="2" to-layer="246" to-port="0" />
-		<edge from-layer="244" from-port="0" to-layer="245" to-port="0" />
-		<edge from-layer="245" from-port="1" to-layer="246" to-port="1" />
+		<edge from-layer="234" from-port="2" to-layer="236" to-port="0" />
+		<edge from-layer="235" from-port="0" to-layer="236" to-port="1" />
+		<edge from-layer="236" from-port="2" to-layer="252" to-port="0" />
+		<edge from-layer="236" from-port="2" to-layer="238" to-port="0" />
+		<edge from-layer="236" from-port="2" to-layer="246" to-port="0" />
+		<edge from-layer="237" from-port="0" to-layer="238" to-port="1" />
+		<edge from-layer="238" from-port="2" to-layer="240" to-port="0" />
+		<edge from-layer="239" from-port="0" to-layer="240" to-port="1" />
+		<edge from-layer="240" from-port="2" to-layer="242" to-port="0" />
+		<edge from-layer="241" from-port="0" to-layer="242" to-port="1" />
+		<edge from-layer="242" from-port="2" to-layer="244" to-port="0" />
+		<edge from-layer="243" from-port="0" to-layer="244" to-port="1" />
+		<edge from-layer="244" from-port="2" to-layer="259" to-port="0" />
+		<edge from-layer="245" from-port="0" to-layer="246" to-port="1" />
 		<edge from-layer="246" from-port="2" to-layer="248" to-port="0" />
 		<edge from-layer="247" from-port="0" to-layer="248" to-port="1" />
 		<edge from-layer="248" from-port="2" to-layer="250" to-port="0" />
 		<edge from-layer="249" from-port="0" to-layer="250" to-port="1" />
-		<edge from-layer="250" from-port="2" to-layer="258" to-port="0" />
-		<edge from-layer="251" from-port="0" to-layer="252" to-port="0" />
-		<edge from-layer="252" from-port="1" to-layer="253" to-port="1" />
-		<edge from-layer="253" from-port="2" to-layer="255" to-port="0" />
-		<edge from-layer="254" from-port="0" to-layer="255" to-port="1" />
-		<edge from-layer="255" from-port="2" to-layer="257" to-port="0" />
-		<edge from-layer="256" from-port="0" to-layer="257" to-port="1" />
-		<edge from-layer="257" from-port="2" to-layer="258" to-port="1" />
-		<edge from-layer="258" from-port="2" to-layer="259" to-port="0" />
-		<edge from-layer="259" from-port="1" to-layer="270" to-port="0" />
-		<edge from-layer="260" from-port="0" to-layer="261" to-port="0" />
-		<edge from-layer="261" from-port="1" to-layer="262" to-port="1" />
-		<edge from-layer="262" from-port="2" to-layer="265" to-port="0" />
-		<edge from-layer="263" from-port="0" to-layer="264" to-port="0" />
-		<edge from-layer="264" from-port="1" to-layer="265" to-port="1" />
+		<edge from-layer="250" from-port="2" to-layer="259" to-port="1" />
+		<edge from-layer="251" from-port="0" to-layer="252" to-port="1" />
+		<edge from-layer="252" from-port="2" to-layer="254" to-port="0" />
+		<edge from-layer="253" from-port="0" to-layer="254" to-port="1" />
+		<edge from-layer="254" from-port="2" to-layer="256" to-port="0" />
+		<edge from-layer="255" from-port="0" to-layer="256" to-port="1" />
+		<edge from-layer="256" from-port="2" to-layer="258" 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="259" to-port="2" />
+		<edge from-layer="259" from-port="3" to-layer="261" to-port="0" />
+		<edge from-layer="260" from-port="0" to-layer="261" to-port="1" />
+		<edge from-layer="261" from-port="2" to-layer="263" to-port="0" />
+		<edge from-layer="262" from-port="0" to-layer="263" to-port="1" />
+		<edge from-layer="263" from-port="2" to-layer="265" to-port="0" />
+		<edge from-layer="264" from-port="0" to-layer="265" to-port="1" />
 		<edge from-layer="265" from-port="2" to-layer="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="269" to-port="0" />
-		<edge from-layer="268" from-port="0" to-layer="269" to-port="1" />
-		<edge from-layer="269" from-port="2" to-layer="270" 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="284" to-port="0" />
+		<edge from-layer="268" from-port="2" to-layer="270" to-port="0" />
+		<edge from-layer="269" from-port="0" to-layer="270" to-port="1" />
 		<edge from-layer="270" from-port="2" to-layer="272" to-port="0" />
 		<edge from-layer="271" from-port="0" to-layer="272" to-port="1" />
 		<edge from-layer="272" from-port="2" to-layer="274" to-port="0" />
 		<edge from-layer="273" from-port="0" to-layer="274" to-port="1" />
-		<edge from-layer="274" from-port="2" to-layer="277" to-port="0" />
-		<edge from-layer="275" from-port="0" 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="280" to-port="0" />
-		<edge from-layer="278" from-port="0" to-layer="279" to-port="0" />
-		<edge from-layer="279" from-port="1" to-layer="280" to-port="1" />
-		<edge from-layer="280" from-port="2" to-layer="281" to-port="1" />
+		<edge from-layer="274" from-port="2" to-layer="276" to-port="0" />
+		<edge from-layer="275" from-port="0" to-layer="276" to-port="1" />
+		<edge from-layer="276" from-port="2" to-layer="278" to-port="0" />
+		<edge from-layer="277" from-port="0" to-layer="278" to-port="1" />
+		<edge from-layer="278" from-port="2" to-layer="279" to-port="0" />
+		<edge from-layer="279" from-port="1" to-layer="281" to-port="0" />
+		<edge from-layer="280" from-port="0" to-layer="281" to-port="1" />
 		<edge from-layer="281" from-port="2" to-layer="283" to-port="0" />
-		<edge from-layer="281" from-port="2" to-layer="303" to-port="0" />
 		<edge from-layer="282" from-port="0" to-layer="283" to-port="1" />
-		<edge from-layer="283" from-port="2" to-layer="286" to-port="0" />
-		<edge from-layer="284" from-port="0" to-layer="285" to-port="0" />
-		<edge from-layer="285" from-port="1" to-layer="286" to-port="1" />
-		<edge from-layer="286" from-port="2" to-layer="289" to-port="0" />
-		<edge from-layer="287" from-port="0" to-layer="288" to-port="0" />
-		<edge from-layer="288" from-port="1" to-layer="289" to-port="1" />
-		<edge from-layer="289" from-port="2" to-layer="292" to-port="0" />
-		<edge from-layer="290" from-port="0" to-layer="291" to-port="0" />
-		<edge from-layer="291" from-port="1" to-layer="292" to-port="1" />
-		<edge from-layer="292" from-port="2" to-layer="295" to-port="0" />
-		<edge from-layer="293" from-port="0" to-layer="294" to-port="0" />
-		<edge from-layer="294" from-port="1" to-layer="295" to-port="1" />
-		<edge from-layer="295" from-port="2" to-layer="296" to-port="0" />
-		<edge from-layer="296" from-port="1" to-layer="299" to-port="0" />
-		<edge from-layer="297" from-port="0" to-layer="298" to-port="0" />
-		<edge from-layer="298" from-port="1" to-layer="299" to-port="1" />
-		<edge from-layer="299" from-port="2" to-layer="302" to-port="0" />
-		<edge from-layer="300" from-port="0" to-layer="301" to-port="0" />
-		<edge from-layer="301" from-port="1" to-layer="302" to-port="1" />
-		<edge from-layer="302" from-port="2" to-layer="303" to-port="1" />
-		<edge from-layer="303" from-port="2" to-layer="352" to-port="0" />
-		<edge from-layer="303" from-port="2" to-layer="305" to-port="0" />
-		<edge from-layer="304" from-port="0" to-layer="305" to-port="1" />
-		<edge from-layer="305" from-port="2" to-layer="308" to-port="0" />
-		<edge from-layer="306" from-port="0" to-layer="307" to-port="0" />
-		<edge from-layer="307" from-port="1" to-layer="308" to-port="1" />
-		<edge from-layer="308" from-port="2" to-layer="311" to-port="0" />
-		<edge from-layer="309" from-port="0" to-layer="310" to-port="0" />
-		<edge from-layer="310" from-port="1" to-layer="311" to-port="1" />
-		<edge from-layer="311" from-port="2" to-layer="314" to-port="0" />
-		<edge from-layer="311" from-port="2" to-layer="333" to-port="0" />
-		<edge from-layer="311" from-port="2" to-layer="324" to-port="0" />
-		<edge from-layer="312" from-port="0" to-layer="313" to-port="0" />
-		<edge from-layer="313" from-port="1" to-layer="314" to-port="1" />
-		<edge from-layer="314" from-port="2" to-layer="317" to-port="0" />
-		<edge from-layer="315" from-port="0" to-layer="316" to-port="0" />
-		<edge from-layer="316" from-port="1" to-layer="317" to-port="1" />
+		<edge from-layer="283" from-port="2" to-layer="284" to-port="1" />
+		<edge from-layer="284" from-port="2" to-layer="286" to-port="0" />
+		<edge from-layer="284" from-port="2" to-layer="322" to-port="0" />
+		<edge from-layer="285" from-port="0" to-layer="286" to-port="1" />
+		<edge from-layer="286" from-port="2" to-layer="288" to-port="0" />
+		<edge from-layer="287" from-port="0" to-layer="288" to-port="1" />
+		<edge from-layer="288" from-port="2" to-layer="290" to-port="0" />
+		<edge from-layer="289" from-port="0" to-layer="290" to-port="1" />
+		<edge from-layer="290" from-port="2" to-layer="306" to-port="0" />
+		<edge from-layer="290" from-port="2" to-layer="300" to-port="0" />
+		<edge from-layer="290" from-port="2" to-layer="292" to-port="0" />
+		<edge from-layer="291" from-port="0" to-layer="292" to-port="1" />
+		<edge from-layer="292" from-port="2" to-layer="294" 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="313" to-port="0" />
+		<edge from-layer="299" from-port="0" to-layer="300" to-port="1" />
+		<edge from-layer="300" from-port="2" to-layer="302" to-port="0" />
+		<edge from-layer="301" from-port="0" to-layer="302" to-port="1" />
+		<edge from-layer="302" from-port="2" to-layer="304" to-port="0" />
+		<edge from-layer="303" from-port="0" to-layer="304" to-port="1" />
+		<edge from-layer="304" from-port="2" to-layer="313" to-port="1" />
+		<edge from-layer="305" from-port="0" to-layer="306" to-port="1" />
+		<edge from-layer="306" from-port="2" to-layer="308" to-port="0" />
+		<edge from-layer="307" from-port="0" to-layer="308" to-port="1" />
+		<edge from-layer="308" from-port="2" to-layer="310" to-port="0" />
+		<edge from-layer="309" from-port="0" to-layer="310" to-port="1" />
+		<edge from-layer="310" from-port="2" to-layer="312" to-port="0" />
+		<edge from-layer="311" from-port="0" to-layer="312" to-port="1" />
+		<edge from-layer="312" from-port="2" to-layer="313" to-port="2" />
+		<edge from-layer="313" from-port="3" to-layer="315" to-port="0" />
+		<edge from-layer="314" from-port="0" to-layer="315" to-port="1" />
+		<edge from-layer="315" from-port="2" to-layer="317" to-port="0" />
+		<edge from-layer="316" from-port="0" to-layer="317" to-port="1" />
 		<edge from-layer="317" from-port="2" to-layer="319" to-port="0" />
 		<edge from-layer="318" from-port="0" 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="329" to-port="0" />
-		<edge from-layer="322" from-port="0" to-layer="323" to-port="0" />
-		<edge from-layer="323" from-port="1" to-layer="324" to-port="1" />
+		<edge from-layer="321" from-port="2" to-layer="322" to-port="1" />
+		<edge from-layer="322" from-port="2" to-layer="324" to-port="0" />
+		<edge from-layer="322" from-port="2" to-layer="338" to-port="0" />
+		<edge from-layer="323" from-port="0" to-layer="324" to-port="1" />
 		<edge from-layer="324" from-port="2" to-layer="326" to-port="0" />
 		<edge from-layer="325" from-port="0" to-layer="326" to-port="1" />
 		<edge from-layer="326" from-port="2" to-layer="328" to-port="0" />
 		<edge from-layer="327" from-port="0" to-layer="328" to-port="1" />
-		<edge from-layer="328" from-port="2" to-layer="329" to-port="1" />
-		<edge from-layer="329" from-port="2" to-layer="330" to-port="0" />
-		<edge from-layer="330" from-port="1" to-layer="341" to-port="0" />
-		<edge from-layer="331" from-port="0" to-layer="332" to-port="0" />
-		<edge from-layer="332" from-port="1" to-layer="333" to-port="1" />
-		<edge from-layer="333" from-port="2" to-layer="336" to-port="0" />
-		<edge from-layer="334" from-port="0" to-layer="335" to-port="0" />
-		<edge from-layer="335" from-port="1" to-layer="336" to-port="1" />
-		<edge from-layer="336" from-port="2" to-layer="338" to-port="0" />
-		<edge from-layer="337" from-port="0" to-layer="338" to-port="1" />
+		<edge from-layer="328" from-port="2" to-layer="330" to-port="0" />
+		<edge from-layer="329" from-port="0" to-layer="330" to-port="1" />
+		<edge from-layer="330" from-port="2" to-layer="332" to-port="0" />
+		<edge from-layer="331" from-port="0" to-layer="332" to-port="1" />
+		<edge from-layer="332" from-port="2" to-layer="333" to-port="0" />
+		<edge from-layer="333" from-port="1" to-layer="335" to-port="0" />
+		<edge from-layer="334" from-port="0" to-layer="335" to-port="1" />
+		<edge from-layer="335" from-port="2" to-layer="337" to-port="0" />
+		<edge from-layer="336" from-port="0" to-layer="337" to-port="1" />
+		<edge from-layer="337" from-port="2" to-layer="338" to-port="1" />
 		<edge from-layer="338" from-port="2" to-layer="340" to-port="0" />
 		<edge from-layer="339" from-port="0" to-layer="340" to-port="1" />
-		<edge from-layer="340" from-port="2" to-layer="341" to-port="1" />
-		<edge from-layer="341" from-port="2" to-layer="343" to-port="0" />
-		<edge from-layer="342" from-port="0" to-layer="343" to-port="1" />
-		<edge from-layer="343" from-port="2" to-layer="345" to-port="0" />
-		<edge from-layer="344" from-port="0" to-layer="345" to-port="1" />
-		<edge from-layer="345" from-port="2" to-layer="348" to-port="0" />
-		<edge from-layer="346" from-port="0" to-layer="347" to-port="0" />
-		<edge from-layer="347" from-port="1" to-layer="348" to-port="1" />
-		<edge from-layer="348" from-port="2" to-layer="351" to-port="0" />
-		<edge from-layer="349" from-port="0" to-layer="350" to-port="0" />
-		<edge from-layer="350" from-port="1" to-layer="351" to-port="1" />
-		<edge from-layer="351" from-port="2" to-layer="352" to-port="1" />
-		<edge from-layer="352" from-port="2" to-layer="374" to-port="0" />
-		<edge from-layer="352" from-port="2" to-layer="354" to-port="0" />
-		<edge from-layer="353" from-port="0" to-layer="354" to-port="1" />
-		<edge from-layer="354" from-port="2" to-layer="357" to-port="0" />
-		<edge from-layer="355" from-port="0" to-layer="356" to-port="0" />
-		<edge from-layer="356" from-port="1" to-layer="357" to-port="1" />
-		<edge from-layer="357" from-port="2" to-layer="360" to-port="0" />
-		<edge from-layer="358" from-port="0" to-layer="359" to-port="0" />
-		<edge from-layer="359" from-port="1" to-layer="360" to-port="1" />
-		<edge from-layer="360" from-port="2" to-layer="363" 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="366" to-port="0" />
-		<edge from-layer="364" from-port="0" to-layer="365" to-port="0" />
-		<edge from-layer="365" from-port="1" to-layer="366" to-port="1" />
-		<edge from-layer="366" from-port="2" to-layer="367" to-port="0" />
-		<edge from-layer="367" from-port="1" to-layer="370" to-port="0" />
-		<edge from-layer="368" from-port="0" to-layer="369" to-port="0" />
-		<edge from-layer="369" from-port="1" to-layer="370" to-port="1" />
-		<edge from-layer="370" from-port="2" to-layer="373" to-port="0" />
-		<edge from-layer="371" from-port="0" to-layer="372" to-port="0" />
-		<edge from-layer="372" from-port="1" to-layer="373" to-port="1" />
-		<edge from-layer="373" from-port="2" to-layer="374" to-port="1" />
-		<edge from-layer="374" from-port="2" to-layer="423" to-port="0" />
-		<edge from-layer="374" from-port="2" to-layer="376" to-port="0" />
-		<edge from-layer="375" from-port="0" to-layer="376" to-port="1" />
-		<edge from-layer="376" from-port="2" to-layer="379" to-port="0" />
-		<edge from-layer="377" from-port="0" to-layer="378" to-port="0" />
-		<edge from-layer="378" from-port="1" to-layer="379" to-port="1" />
-		<edge from-layer="379" from-port="2" to-layer="382" to-port="0" />
-		<edge from-layer="380" from-port="0" to-layer="381" to-port="0" />
-		<edge from-layer="381" from-port="1" to-layer="382" to-port="1" />
-		<edge from-layer="382" from-port="2" to-layer="404" to-port="0" />
-		<edge from-layer="382" from-port="2" to-layer="395" to-port="0" />
-		<edge from-layer="382" from-port="2" 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="388" to-port="0" />
-		<edge from-layer="386" from-port="0" to-layer="387" to-port="0" />
-		<edge from-layer="387" from-port="1" to-layer="388" to-port="1" />
-		<edge from-layer="388" from-port="2" to-layer="390" to-port="0" />
-		<edge from-layer="389" from-port="0" to-layer="390" to-port="1" />
-		<edge from-layer="390" from-port="2" to-layer="392" to-port="0" />
-		<edge from-layer="391" from-port="0" to-layer="392" to-port="1" />
-		<edge from-layer="392" from-port="2" to-layer="400" to-port="0" />
-		<edge from-layer="393" from-port="0" to-layer="394" to-port="0" />
-		<edge from-layer="394" from-port="1" to-layer="395" to-port="1" />
-		<edge from-layer="395" from-port="2" to-layer="397" to-port="0" />
-		<edge from-layer="396" from-port="0" to-layer="397" to-port="1" />
-		<edge from-layer="397" from-port="2" to-layer="399" to-port="0" />
-		<edge from-layer="398" from-port="0" to-layer="399" to-port="1" />
-		<edge from-layer="399" from-port="2" to-layer="400" to-port="1" />
-		<edge from-layer="400" from-port="2" to-layer="401" to-port="0" />
-		<edge from-layer="401" from-port="1" to-layer="412" to-port="0" />
-		<edge from-layer="402" from-port="0" to-layer="403" to-port="0" />
-		<edge from-layer="403" from-port="1" to-layer="404" to-port="1" />
-		<edge from-layer="404" from-port="2" to-layer="407" to-port="0" />
-		<edge from-layer="405" from-port="0" to-layer="406" to-port="0" />
-		<edge from-layer="406" from-port="1" to-layer="407" to-port="1" />
-		<edge from-layer="407" from-port="2" to-layer="409" to-port="0" />
-		<edge from-layer="408" from-port="0" to-layer="409" to-port="1" />
-		<edge from-layer="409" from-port="2" to-layer="411" to-port="0" />
-		<edge from-layer="410" from-port="0" to-layer="411" to-port="1" />
-		<edge from-layer="411" from-port="2" 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="416" to-port="0" />
-		<edge from-layer="415" from-port="0" to-layer="416" to-port="1" />
-		<edge from-layer="416" from-port="2" to-layer="419" to-port="0" />
-		<edge from-layer="417" from-port="0" to-layer="418" to-port="0" />
-		<edge from-layer="418" from-port="1" to-layer="419" to-port="1" />
-		<edge from-layer="419" from-port="2" to-layer="422" to-port="0" />
-		<edge from-layer="420" from-port="0" to-layer="421" to-port="0" />
-		<edge from-layer="421" from-port="1" to-layer="422" to-port="1" />
-		<edge from-layer="422" from-port="2" to-layer="423" to-port="1" />
-		<edge from-layer="423" from-port="2" to-layer="425" to-port="0" />
-		<edge from-layer="423" from-port="2" to-layer="445" to-port="0" />
-		<edge from-layer="424" from-port="0" to-layer="425" to-port="1" />
-		<edge from-layer="425" from-port="2" to-layer="428" to-port="0" />
-		<edge from-layer="426" from-port="0" to-layer="427" to-port="0" />
-		<edge from-layer="427" from-port="1" to-layer="428" to-port="1" />
-		<edge from-layer="428" from-port="2" to-layer="431" to-port="0" />
-		<edge from-layer="429" from-port="0" to-layer="430" to-port="0" />
-		<edge from-layer="430" from-port="1" to-layer="431" to-port="1" />
-		<edge from-layer="431" from-port="2" to-layer="434" to-port="0" />
-		<edge from-layer="432" from-port="0" to-layer="433" to-port="0" />
-		<edge from-layer="433" from-port="1" to-layer="434" to-port="1" />
-		<edge from-layer="434" from-port="2" to-layer="437" to-port="0" />
-		<edge from-layer="435" from-port="0" to-layer="436" to-port="0" />
-		<edge from-layer="436" from-port="1" to-layer="437" to-port="1" />
-		<edge from-layer="437" from-port="2" to-layer="438" to-port="0" />
-		<edge from-layer="438" from-port="1" to-layer="441" to-port="0" />
-		<edge from-layer="439" from-port="0" 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="444" to-port="0" />
-		<edge from-layer="442" from-port="0" to-layer="443" to-port="0" />
-		<edge from-layer="443" from-port="1" 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="447" to-port="0" />
-		<edge from-layer="446" from-port="0" to-layer="447" to-port="1" />
-		<edge from-layer="447" from-port="2" to-layer="450" to-port="0" />
-		<edge from-layer="448" from-port="0" to-layer="449" to-port="0" />
-		<edge from-layer="449" from-port="1" to-layer="450" to-port="1" />
-		<edge from-layer="450" from-port="2" to-layer="453" to-port="0" />
-		<edge from-layer="451" from-port="0" to-layer="452" to-port="0" />
-		<edge from-layer="452" from-port="1" to-layer="453" to-port="1" />
-		<edge from-layer="453" from-port="2" to-layer="454" to-port="0" />
+		<edge from-layer="340" from-port="2" to-layer="342" to-port="0" />
+		<edge from-layer="341" from-port="0" to-layer="342" to-port="1" />
+		<edge from-layer="342" from-port="2" to-layer="344" to-port="0" />
+		<edge from-layer="343" from-port="0" to-layer="344" to-port="1" />
+		<edge from-layer="344" from-port="2" to-layer="345" to-port="0" />
 	</edges>
 	<rt_info>
-		<Runtime_version value="2024.5.0-17285-ea5c1dcfdf9-releases/2024/5" />
+		<Runtime_version value="2025.2.0-19120-87425bc78ca-releases/2025/2" />
 		<conversion_parameters>
 			<framework value="pytorch" />
 			<is_python_object value="True" />
 		</conversion_parameters>
 		<optimum>
-			<optimum_intel_version value="1.21.0.dev0+d357376" />
-			<optimum_version value="1.23.3" />
-			<pytorch_version value="2.5.1" />
-			<transformers_version value="4.46.3" />
+			<optimum_intel_version value="1.24.0.dev0+08e3008" />
+			<optimum_version value="1.25.3" />
+			<pytorch_version value="2.7.0+cpu" />
+			<transformers_version value="4.51.3" />
 		</optimum>
 	</rt_info>
 </net>