File size: 23,129 Bytes
9422383 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 | // Fixed top-down neck partitions. Reserved values reproduce XNNPACK's
// performance-sensitive declaration inventory but are not part of data flow.
// Projects the backbone P4 feature and adds the upsampled P5 projection to form
// the model.13 fusion input.
bool BuildNeckP4Fusion(const GraphContext& context) {
FixedGraphBuilder graph(context, GraphId::kNeckP4Fusion, NeckP4FusionBindings::kCount);
if (!graph.Begin()) return false;
const GraphValue upsampled_p5_input = graph.DefineInput(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)},
NeckP4FusionBindings::kUpsampledP5Input);
const GraphValue backbone_p4_input = graph.DefineInput(
ValueType::kFp16, {Constant(1), Constant(128), ResolutionDivided(16), ResolutionDivided(16)},
NeckP4FusionBindings::kBackboneP4Input);
[[maybe_unused]] const GraphValue reserved_value_2 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue transpose_1_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue model13_cv1_conv_weight_input256384 =
graph.DefineWeight(ValueType::kFp32, {Constant(128), Constant(1), Constant(1), Constant(128)},
WeightSlot::kModel13Cv1ConvWeightInput256384);
[[maybe_unused]] const GraphValue reserved_value_5 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue model13_cv1_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(128)}, WeightSlot::kModel13Cv1ConvBias);
[[maybe_unused]] const GraphValue reserved_value_7 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue model13_cv1_conv_weight_input256384_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue fusion_output = graph.DefineOutput(
ValueType::kFp16, {Constant(1), Constant(128), ResolutionDivided(16), ResolutionDivided(16)},
NeckP4FusionBindings::kFusionOutput);
const GraphValue residual_sum_1 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue residual_sum_1_activation = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
if (!graph.DefinitionsValid()) return false;
if (!graph.Transpose(backbone_p4_input, transpose_1_output,
{Constant(0), Constant(2), Constant(3), Constant(1)}, 0))
return false;
if (!graph.Convolution(transpose_1_output, model13_cv1_conv_weight_input256384,
model13_cv1_conv_bias, model13_cv1_conv_weight_input256384_output,
Padding{Constant(0), Constant(0), Constant(0), Constant(0)},
Kernel{Constant(1), Constant(1)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)},
Groups{Constant(1), Constant(128), Constant(128)}, 0))
return false;
if (!graph.Add(upsampled_p5_input, model13_cv1_conv_weight_input256384_output, residual_sum_1, 0))
return false;
if (!graph.Silu(residual_sum_1, residual_sum_1_activation, 0)) return false;
if (!graph.Transpose(residual_sum_1_activation, fusion_output,
{Constant(0), Constant(3), Constant(1), Constant(2)}, 0))
return false;
return graph.Finish();
}
// Runs the model.13 C2f block and emits both the fused P4 feature and the
// narrower projection that the schedule upsamples toward P3.
bool BuildNeckP4Block(const GraphContext& context) {
FixedGraphBuilder graph(context, GraphId::kNeckP4Block, NeckP4BlockBindings::kCount);
if (!graph.Begin()) return false;
const GraphValue main_input = graph.DefineInput(
ValueType::kFp16, {Constant(1), Constant(64), ResolutionDivided(16), ResolutionDivided(16)},
NeckP4BlockBindings::kMainInput);
const GraphValue transpose_1_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(64)});
const GraphValue skip_input = graph.DefineInput(
ValueType::kFp16, {Constant(1), Constant(64), ResolutionDivided(16), ResolutionDivided(16)},
NeckP4BlockBindings::kSkipInput);
const GraphValue transpose_2_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(64)});
const GraphValue model13_m0_cv1_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(32), Constant(1), Constant(1), Constant(64)},
WeightSlot::kModel13M0Cv1ConvWeight);
[[maybe_unused]] const GraphValue reserved_value_5 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_cv1_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(32)}, WeightSlot::kModel13M0Cv1ConvBias);
const GraphValue model13_m0_cv2_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(32), Constant(1), Constant(1), Constant(64)},
WeightSlot::kModel13M0Cv2ConvWeight);
[[maybe_unused]] const GraphValue reserved_value_8 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_cv2_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(32)}, WeightSlot::kModel13M0Cv2ConvBias);
[[maybe_unused]] const GraphValue reserved_value_10 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
[[maybe_unused]] const GraphValue reserved_value_11 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_cv1_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_cv2_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m0_cv1_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(32), Constant(3), Constant(3), Constant(32)},
WeightSlot::kModel13M0M0Cv1ConvWeight);
const GraphValue model13_m0_m0_cv1_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m0_cv1_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(32)}, WeightSlot::kModel13M0M0Cv1ConvBias);
[[maybe_unused]] const GraphValue reserved_value_17 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m0_cv1_conv_activation = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m0_cv2_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(32), Constant(3), Constant(3), Constant(32)},
WeightSlot::kModel13M0M0Cv2ConvWeight);
const GraphValue model13_m0_m0_cv2_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m0_cv2_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(32)}, WeightSlot::kModel13M0M0Cv2ConvBias);
[[maybe_unused]] const GraphValue reserved_value_22 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m0_cv2_conv_activation = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue residual_sum_1 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m1_cv1_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(32), Constant(3), Constant(3), Constant(32)},
WeightSlot::kModel13M0M1Cv1ConvWeight);
const GraphValue model13_m0_m1_cv1_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m1_cv1_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(32)}, WeightSlot::kModel13M0M1Cv1ConvBias);
[[maybe_unused]] const GraphValue reserved_value_28 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m1_cv1_conv_activation = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m1_cv2_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(32), Constant(3), Constant(3), Constant(32)},
WeightSlot::kModel13M0M1Cv2ConvWeight);
const GraphValue model13_m0_m1_cv2_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m1_cv2_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(32)}, WeightSlot::kModel13M0M1Cv2ConvBias);
[[maybe_unused]] const GraphValue reserved_value_33 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue model13_m0_m1_cv2_conv_activation = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue residual_sum_2 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(32)});
const GraphValue concatenation_1 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(64)});
const GraphValue model13_m0_cv3_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(64), Constant(1), Constant(1), Constant(64)},
WeightSlot::kModel13M0Cv3ConvWeight);
[[maybe_unused]] const GraphValue reserved_value_38 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(64)});
const GraphValue model13_m0_cv3_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(64)}, WeightSlot::kModel13M0Cv3ConvBias);
[[maybe_unused]] const GraphValue reserved_value_40 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(64)});
const GraphValue model13_m0_cv3_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(64)});
const GraphValue concatenation_2 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(192)});
const GraphValue model13_cv2_conv_weight =
graph.DefineWeight(ValueType::kFp32, {Constant(128), Constant(1), Constant(1), Constant(192)},
WeightSlot::kModel13Cv2ConvWeight);
[[maybe_unused]] const GraphValue reserved_value_44 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue model13_cv2_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(128)}, WeightSlot::kModel13Cv2ConvBias);
[[maybe_unused]] const GraphValue reserved_value_46 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue model13_cv2_conv_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(128)});
const GraphValue neck_p4_output = graph.DefineOutput(
ValueType::kFp16, {Constant(1), Constant(128), ResolutionDivided(16), ResolutionDivided(16)},
NeckP4BlockBindings::kNeckP4Output);
const GraphValue model16_cv1_conv_weight_input0128 =
graph.DefineWeight(ValueType::kFp32, {Constant(64), Constant(1), Constant(1), Constant(128)},
WeightSlot::kModel16Cv1ConvWeightInput0128);
const GraphValue projection_zero_bias64 =
graph.DefineWeight(ValueType::kFp32, {Constant(64)}, WeightSlot::kProjectionZeroBias64);
const GraphValue p4_projection_output = graph.DefineOutput(
ValueType::kFp16, {Constant(1), ResolutionDivided(16), ResolutionDivided(16), Constant(64)},
NeckP4BlockBindings::kP4ProjectionOutput);
[[maybe_unused]] const GraphValue reserved_value_52 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), Constant(64), ResolutionDivided(16), ResolutionDivided(16)});
if (!graph.DefinitionsValid()) return false;
if (!graph.Transpose(main_input, transpose_1_output,
{Constant(0), Constant(2), Constant(3), Constant(1)}, 0))
return false;
if (!graph.Transpose(skip_input, transpose_2_output,
{Constant(0), Constant(2), Constant(3), Constant(1)}, 0))
return false;
if (!graph.Convolution(
transpose_1_output, model13_m0_cv1_conv_weight, model13_m0_cv1_conv_bias,
model13_m0_cv1_conv_output, Padding{Constant(0), Constant(0), Constant(0), Constant(0)},
Kernel{Constant(1), Constant(1)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)}, Groups{Constant(1), Constant(64), Constant(32)},
XNN_FLAG_YOLO26_GEMM_SILU_FINAL_STORE))
return false;
if (!graph.Convolution(
transpose_1_output, model13_m0_cv2_conv_weight, model13_m0_cv2_conv_bias,
model13_m0_cv2_conv_output, Padding{Constant(0), Constant(0), Constant(0), Constant(0)},
Kernel{Constant(1), Constant(1)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)}, Groups{Constant(1), Constant(64), Constant(32)},
XNN_FLAG_YOLO26_GEMM_SILU_FINAL_STORE))
return false;
if (!graph.Convolution(model13_m0_cv1_conv_output, model13_m0_m0_cv1_conv_weight,
model13_m0_m0_cv1_conv_bias, model13_m0_m0_cv1_conv_output,
Padding{Constant(1), Constant(1), Constant(1), Constant(1)},
Kernel{Constant(3), Constant(3)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)},
Groups{Constant(1), Constant(32), Constant(32)}, 0))
return false;
if (!graph.Silu(model13_m0_m0_cv1_conv_output, model13_m0_m0_cv1_conv_activation, 0))
return false;
if (!graph.Convolution(model13_m0_m0_cv1_conv_activation, model13_m0_m0_cv2_conv_weight,
model13_m0_m0_cv2_conv_bias, model13_m0_m0_cv2_conv_output,
Padding{Constant(1), Constant(1), Constant(1), Constant(1)},
Kernel{Constant(3), Constant(3)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)},
Groups{Constant(1), Constant(32), Constant(32)}, 0))
return false;
if (!graph.Silu(model13_m0_m0_cv2_conv_output, model13_m0_m0_cv2_conv_activation, 0))
return false;
if (!graph.Add(model13_m0_cv1_conv_output, model13_m0_m0_cv2_conv_activation, residual_sum_1, 0))
return false;
if (!graph.Convolution(residual_sum_1, model13_m0_m1_cv1_conv_weight, model13_m0_m1_cv1_conv_bias,
model13_m0_m1_cv1_conv_output,
Padding{Constant(1), Constant(1), Constant(1), Constant(1)},
Kernel{Constant(3), Constant(3)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)},
Groups{Constant(1), Constant(32), Constant(32)}, 0))
return false;
if (!graph.Silu(model13_m0_m1_cv1_conv_output, model13_m0_m1_cv1_conv_activation, 0))
return false;
if (!graph.Convolution(model13_m0_m1_cv1_conv_activation, model13_m0_m1_cv2_conv_weight,
model13_m0_m1_cv2_conv_bias, model13_m0_m1_cv2_conv_output,
Padding{Constant(1), Constant(1), Constant(1), Constant(1)},
Kernel{Constant(3), Constant(3)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)},
Groups{Constant(1), Constant(32), Constant(32)}, 0))
return false;
if (!graph.Silu(model13_m0_m1_cv2_conv_output, model13_m0_m1_cv2_conv_activation, 0))
return false;
if (!graph.Add(residual_sum_1, model13_m0_m1_cv2_conv_activation, residual_sum_2, 0))
return false;
if (!graph.Concatenate({residual_sum_2, model13_m0_cv2_conv_output}, concatenation_1, Constant(3),
0))
return false;
if (!graph.Convolution(
concatenation_1, model13_m0_cv3_conv_weight, model13_m0_cv3_conv_bias,
model13_m0_cv3_conv_output, Padding{Constant(0), Constant(0), Constant(0), Constant(0)},
Kernel{Constant(1), Constant(1)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)}, Groups{Constant(1), Constant(64), Constant(64)},
XNN_FLAG_YOLO26_GEMM_SILU_FINAL_STORE))
return false;
if (!graph.Concatenate({transpose_2_output, transpose_1_output, model13_m0_cv3_conv_output},
concatenation_2, Constant(3), 0))
return false;
if (!graph.Convolution(
concatenation_2, model13_cv2_conv_weight, model13_cv2_conv_bias, model13_cv2_conv_output,
Padding{Constant(0), Constant(0), Constant(0), Constant(0)},
Kernel{Constant(1), Constant(1)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)}, Groups{Constant(1), Constant(192), Constant(128)},
XNN_FLAG_YOLO26_GEMM_SILU_FINAL_STORE))
return false;
if (!graph.Transpose(model13_cv2_conv_output, neck_p4_output,
{Constant(0), Constant(3), Constant(1), Constant(2)}, 0))
return false;
if (!graph.Convolution(
model13_cv2_conv_output, model16_cv1_conv_weight_input0128, projection_zero_bias64,
p4_projection_output, Padding{Constant(0), Constant(0), Constant(0), Constant(0)},
Kernel{Constant(1), Constant(1)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)}, Groups{Constant(1), Constant(128), Constant(64)}, 0))
return false;
return graph.Finish();
}
// Projects the backbone P3 feature and adds the upsampled P4 projection. The
// schedule splits the result into the two model.16 branches.
bool BuildNeckP3Fusion(const GraphContext& context) {
FixedGraphBuilder graph(context, GraphId::kNeckP3Fusion, NeckP3FusionBindings::kCount);
if (!graph.Begin()) return false;
[[maybe_unused]] const GraphValue reserved_value_0 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), Constant(128), ResolutionDivided(8), ResolutionDivided(8)});
[[maybe_unused]] const GraphValue reserved_value_1 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), Constant(128), ResolutionDivided(8), ResolutionDivided(8)});
[[maybe_unused]] const GraphValue reserved_value_2 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), Constant(256), ResolutionDivided(8), ResolutionDivided(8)});
const GraphValue upsampled_p4_input = graph.DefineInput(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(64)},
NeckP3FusionBindings::kUpsampledP4Input);
const GraphValue model16_cv1_conv_weight_input128256 =
graph.DefineWeight(ValueType::kFp32, {Constant(64), Constant(1), Constant(1), Constant(128)},
WeightSlot::kModel16Cv1ConvWeightInput128256);
[[maybe_unused]] const GraphValue reserved_value_5 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(64)});
const GraphValue model16_cv1_conv_bias =
graph.DefineWeight(ValueType::kFp32, {Constant(64)}, WeightSlot::kModel16Cv1ConvBias);
[[maybe_unused]] const GraphValue reserved_value_7 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(64)});
const GraphValue model16_cv1_conv_weight_input128256_output = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(64)});
[[maybe_unused]] const GraphValue reserved_value_9 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), Constant(64), ResolutionDivided(8), ResolutionDivided(8)});
[[maybe_unused]] const GraphValue reserved_value_10 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(64)});
const GraphValue fusion_output = graph.DefineInput(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(128)},
NeckP3FusionBindings::kFusionOutput);
[[maybe_unused]] const GraphValue reserved_value_12 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(128)});
const GraphValue residual_sum_1 = graph.DefineIntermediate(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(64)});
const GraphValue backbone_p3_input = graph.DefineOutput(
ValueType::kFp16, {Constant(1), ResolutionDivided(8), ResolutionDivided(8), Constant(64)},
NeckP3FusionBindings::kBackboneP3Input);
if (!graph.DefinitionsValid()) return false;
if (!graph.Convolution(fusion_output, model16_cv1_conv_weight_input128256, model16_cv1_conv_bias,
model16_cv1_conv_weight_input128256_output,
Padding{Constant(0), Constant(0), Constant(0), Constant(0)},
Kernel{Constant(1), Constant(1)}, Stride{Constant(1), Constant(1)},
Dilation{Constant(1), Constant(1)},
Groups{Constant(1), Constant(128), Constant(64)}, 0))
return false;
if (!graph.Add(upsampled_p4_input, model16_cv1_conv_weight_input128256_output, residual_sum_1, 0))
return false;
if (!graph.Silu(residual_sum_1, backbone_p3_input, 0)) return false;
return graph.Finish();
}
|