使用rknn-toolkit1.6.0转换onnx模型，加载模型错误。

运行信息：
W Verbose file path is invalid, debug info will not dump to file.
--> config model
done
--> Loading model
I Start importing onnx...
W Call onnx.optimizer.optimize fail, skip optimize
I Current ONNX Model use ir_version 6 opset_version 11
D Calc tensor Initializer_480 (5,)
D Calc tensor Initializer_460 (5,)
D Calc tensor Initializer_440 (5,)
D Calc tensor Initializer_374 (4,)
D Calc tensor Initializer_329 (4,)
D Calc tensor Initializer_322 (0,)
D Calc tensor Initializer_178 (1,)
D Calc tensor Initializer_169 (1,)
D Calc tensor Initializer_168 (1,)
D Calc tensor Initializer_167 (1,)
D Calc tensor Slice_171 (1, 3, 320, 640)
D Calc tensor Slice_181 (1, 3, 320, 640)
D Calc tensor Slice_191 (1, 3, 320, 640)
D Calc tensor Slice_201 (1, 3, 320, 640)
D Calc tensor Initializer_model.9.m.0.cv2.conv.weight (256, 256, 3, 3)
D Calc tensor Initializer_model.9.m.0.cv2.conv.bias (256,)
D Calc tensor Initializer_model.9.m.0.cv1.conv.weight (256, 256, 1, 1)
D Calc tensor Initializer_model.9.m.0.cv1.conv.bias (256,)
D Calc tensor Initializer_model.9.cv4.conv.weight (512, 512, 1, 1)
D Calc tensor Initializer_model.9.cv4.conv.bias (512,)
D Calc tensor Initializer_model.9.cv3.weight (256, 256, 1, 1)
D Calc tensor Initializer_model.9.cv2.weight (256, 512, 1, 1)
D Calc tensor Initializer_model.9.cv1.conv.weight (256, 512, 1, 1)
D Calc tensor Initializer_model.9.cv1.conv.bias (256,)
D Calc tensor Initializer_model.9.bn.weight (512,)
D Calc tensor Initializer_model.9.bn.running_var (512,)
D Calc tensor Initializer_model.9.bn.running_mean (512,)
D Calc tensor Initializer_model.9.bn.bias (512,)
D Calc tensor Initializer_model.8.cv2.conv.weight (512, 1024, 1, 1)
D Calc tensor Initializer_model.8.cv2.conv.bias (512,)
D Calc tensor Initializer_model.8.cv1.conv.weight (256, 512, 1, 1)
D Calc tensor Initializer_model.8.cv1.conv.bias (256,)
D Calc tensor Initializer_model.7.conv.weight (512, 256, 3, 3)
D Calc tensor Initializer_model.7.conv.bias (512,)
D Calc tensor Initializer_model.6.m.2.cv2.conv.weight (128, 128, 3, 3)
D Calc tensor Initializer_model.6.m.2.cv2.conv.bias (128,)
D Calc tensor Initializer_model.6.m.2.cv1.conv.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.6.m.2.cv1.conv.bias (128,)
D Calc tensor Initializer_model.6.m.1.cv2.conv.weight (128, 128, 3, 3)
D Calc tensor Initializer_model.6.m.1.cv2.conv.bias (128,)
D Calc tensor Initializer_model.6.m.1.cv1.conv.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.6.m.1.cv1.conv.bias (128,)
D Calc tensor Initializer_model.6.m.0.cv2.conv.weight (128, 128, 3, 3)
D Calc tensor Initializer_model.6.m.0.cv2.conv.bias (128,)
D Calc tensor Initializer_model.6.m.0.cv1.conv.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.6.m.0.cv1.conv.bias (128,)
D Calc tensor Initializer_model.6.cv4.conv.weight (256, 256, 1, 1)
D Calc tensor Initializer_model.6.cv4.conv.bias (256,)
D Calc tensor Initializer_model.6.cv3.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.6.cv2.weight (128, 256, 1, 1)
D Calc tensor Initializer_model.6.cv1.conv.weight (128, 256, 1, 1)
D Calc tensor Initializer_model.6.cv1.conv.bias (128,)
D Calc tensor Initializer_model.6.bn.weight (256,)
D Calc tensor Initializer_model.6.bn.running_var (256,)
D Calc tensor Initializer_model.6.bn.running_mean (256,)
D Calc tensor Initializer_model.6.bn.bias (256,)
D Calc tensor Initializer_model.5.conv.weight (256, 128, 3, 3)
D Calc tensor Initializer_model.5.conv.bias (256,)
D Calc tensor Initializer_model.4.m.2.cv2.conv.weight (64, 64, 3, 3)
D Calc tensor Initializer_model.4.m.2.cv2.conv.bias (64,)
D Calc tensor Initializer_model.4.m.2.cv1.conv.weight (64, 64, 1, 1)
D Calc tensor Initializer_model.4.m.2.cv1.conv.bias (64,)
D Calc tensor Initializer_model.4.m.1.cv2.conv.weight (64, 64, 3, 3)
D Calc tensor Initializer_model.4.m.1.cv2.conv.bias (64,)
D Calc tensor Initializer_model.4.m.1.cv1.conv.weight (64, 64, 1, 1)
D Calc tensor Initializer_model.4.m.1.cv1.conv.bias (64,)
D Calc tensor Initializer_model.4.m.0.cv2.conv.weight (64, 64, 3, 3)
D Calc tensor Initializer_model.4.m.0.cv2.conv.bias (64,)
D Calc tensor Initializer_model.4.m.0.cv1.conv.weight (64, 64, 1, 1)
D Calc tensor Initializer_model.4.m.0.cv1.conv.bias (64,)
D Calc tensor Initializer_model.4.cv4.conv.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.4.cv4.conv.bias (128,)
D Calc tensor Initializer_model.4.cv3.weight (64, 64, 1, 1)
D Calc tensor Initializer_model.4.cv2.weight (64, 128, 1, 1)
D Calc tensor Initializer_model.4.cv1.conv.weight (64, 128, 1, 1)
D Calc tensor Initializer_model.4.cv1.conv.bias (64,)
D Calc tensor Initializer_model.4.bn.weight (128,)
D Calc tensor Initializer_model.4.bn.running_var (128,)
D Calc tensor Initializer_model.4.bn.running_mean (128,)
D Calc tensor Initializer_model.4.bn.bias (128,)
D Calc tensor Initializer_model.3.conv.weight (128, 64, 3, 3)
D Calc tensor Initializer_model.3.conv.bias (128,)
D Calc tensor Initializer_model.24.m.2.weight (24, 512, 1, 1)
D Calc tensor Initializer_model.24.m.2.bias (24,)
D Calc tensor Initializer_model.24.m.1.weight (24, 256, 1, 1)
D Calc tensor Initializer_model.24.m.1.bias (24,)
D Calc tensor Initializer_model.24.m.0.weight (24, 128, 1, 1)
D Calc tensor Initializer_model.24.m.0.bias (24,)
D Calc tensor Initializer_model.23.m.0.cv2.conv.weight (256, 256, 3, 3)
D Calc tensor Initializer_model.23.m.0.cv2.conv.bias (256,)
D Calc tensor Initializer_model.23.m.0.cv1.conv.weight (256, 256, 1, 1)
D Calc tensor Initializer_model.23.m.0.cv1.conv.bias (256,)
D Calc tensor Initializer_model.23.cv4.conv.weight (512, 512, 1, 1)
D Calc tensor Initializer_model.23.cv4.conv.bias (512,)
D Calc tensor Initializer_model.23.cv3.weight (256, 256, 1, 1)
D Calc tensor Initializer_model.23.cv2.weight (256, 512, 1, 1)
D Calc tensor Initializer_model.23.cv1.conv.weight (256, 512, 1, 1)
D Calc tensor Initializer_model.23.cv1.conv.bias (256,)
D Calc tensor Initializer_model.23.bn.weight (512,)
D Calc tensor Initializer_model.23.bn.running_var (512,)
D Calc tensor Initializer_model.23.bn.running_mean (512,)
D Calc tensor Initializer_model.23.bn.bias (512,)
D Calc tensor Initializer_model.21.conv.weight (256, 256, 3, 3)
D Calc tensor Initializer_model.21.conv.bias (256,)
D Calc tensor Initializer_model.20.m.0.cv2.conv.weight (128, 128, 3, 3)
D Calc tensor Initializer_model.20.m.0.cv2.conv.bias (128,)
D Calc tensor Initializer_model.20.m.0.cv1.conv.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.20.m.0.cv1.conv.bias (128,)
D Calc tensor Initializer_model.20.cv4.conv.weight (256, 256, 1, 1)
D Calc tensor Initializer_model.20.cv4.conv.bias (256,)
D Calc tensor Initializer_model.20.cv3.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.20.cv2.weight (128, 256, 1, 1)
D Calc tensor Initializer_model.20.cv1.conv.weight (128, 256, 1, 1)
D Calc tensor Initializer_model.20.cv1.conv.bias (128,)
D Calc tensor Initializer_model.20.bn.weight (256,)
D Calc tensor Initializer_model.20.bn.running_var (256,)
D Calc tensor Initializer_model.20.bn.running_mean (256,)
D Calc tensor Initializer_model.20.bn.bias (256,)
D Calc tensor Initializer_model.2.m.0.cv2.conv.weight (32, 32, 3, 3)
D Calc tensor Initializer_model.2.m.0.cv2.conv.bias (32,)
D Calc tensor Initializer_model.2.m.0.cv1.conv.weight (32, 32, 1, 1)
D Calc tensor Initializer_model.2.m.0.cv1.conv.bias (32,)
D Calc tensor Initializer_model.2.cv4.conv.weight (64, 64, 1, 1)
D Calc tensor Initializer_model.2.cv4.conv.bias (64,)
D Calc tensor Initializer_model.2.cv3.weight (32, 32, 1, 1)
D Calc tensor Initializer_model.2.cv2.weight (32, 64, 1, 1)
D Calc tensor Initializer_model.2.cv1.conv.weight (32, 64, 1, 1)
D Calc tensor Initializer_model.2.cv1.conv.bias (32,)
D Calc tensor Initializer_model.2.bn.weight (64,)
D Calc tensor Initializer_model.2.bn.running_var (64,)
D Calc tensor Initializer_model.2.bn.running_mean (64,)
D Calc tensor Initializer_model.2.bn.bias (64,)
D Calc tensor Initializer_model.18.conv.weight (128, 128, 3, 3)
D Calc tensor Initializer_model.18.conv.bias (128,)
D Calc tensor Initializer_model.17.m.0.cv2.conv.weight (64, 64, 3, 3)
D Calc tensor Initializer_model.17.m.0.cv2.conv.bias (64,)
D Calc tensor Initializer_model.17.m.0.cv1.conv.weight (64, 64, 1, 1)
D Calc tensor Initializer_model.17.m.0.cv1.conv.bias (64,)
D Calc tensor Initializer_model.17.cv4.conv.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.17.cv4.conv.bias (128,)
D Calc tensor Initializer_model.17.cv3.weight (64, 64, 1, 1)
D Calc tensor Initializer_model.17.cv2.weight (64, 256, 1, 1)
D Calc tensor Initializer_model.17.cv1.conv.weight (64, 256, 1, 1)
D Calc tensor Initializer_model.17.cv1.conv.bias (64,)
D Calc tensor Initializer_model.17.bn.weight (128,)
D Calc tensor Initializer_model.17.bn.running_var (128,)
D Calc tensor Initializer_model.17.bn.running_mean (128,)
D Calc tensor Initializer_model.17.bn.bias (128,)
D Calc tensor Initializer_model.14.conv.weight (128, 256, 1, 1)
D Calc tensor Initializer_model.14.conv.bias (128,)
D Calc tensor Initializer_model.13.m.0.cv2.conv.weight (128, 128, 3, 3)
D Calc tensor Initializer_model.13.m.0.cv2.conv.bias (128,)
D Calc tensor Initializer_model.13.m.0.cv1.conv.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.13.m.0.cv1.conv.bias (128,)
D Calc tensor Initializer_model.13.cv4.conv.weight (256, 256, 1, 1)
D Calc tensor Initializer_model.13.cv4.conv.bias (256,)
D Calc tensor Initializer_model.13.cv3.weight (128, 128, 1, 1)
D Calc tensor Initializer_model.13.cv2.weight (128, 512, 1, 1)
D Calc tensor Initializer_model.13.cv1.conv.weight (128, 512, 1, 1)
D Calc tensor Initializer_model.13.cv1.conv.bias (128,)
D Calc tensor Initializer_model.13.bn.weight (256,)
D Calc tensor Initializer_model.13.bn.running_var (256,)
D Calc tensor Initializer_model.13.bn.running_mean (256,)
D Calc tensor Initializer_model.13.bn.bias (256,)
D Calc tensor Initializer_model.10.conv.weight (256, 512, 1, 1)
D Calc tensor Initializer_model.10.conv.bias (256,)
D Calc tensor Initializer_model.1.conv.weight (64, 32, 3, 3)
D Calc tensor Initializer_model.1.conv.bias (64,)
D Calc tensor Initializer_model.0.conv.conv.weight (32, 12, 3, 3)
D Calc tensor Initializer_model.0.conv.conv.bias (32,)
D Calc tensor Initializer_483 (1,)
D Calc tensor Slice_176 (1, 3, 320, 320)
D Calc tensor Slice_186 (1, 3, 320, 320)
D Calc tensor Slice_196 (1, 3, 320, 320)
D Calc tensor Slice_206 (1, 3, 320, 320)
D Calc tensor Concat_207 (1, 12, 320, 320)
D Calc tensor Conv_208 (1, 32, 320, 320)
D Calc tensor LeakyRelu_209 (1, 32, 320, 320)
D Calc tensor Conv_210 (1, 64, 160, 160)
D Calc tensor LeakyRelu_211 (1, 64, 160, 160)
D Calc tensor Conv_212 (1, 32, 160, 160)
D Calc tensor LeakyRelu_213 (1, 32, 160, 160)
D Calc tensor Conv_214 (1, 32, 160, 160)
D Calc tensor LeakyRelu_215 (1, 32, 160, 160)
D Calc tensor Conv_216 (1, 32, 160, 160)
D Calc tensor LeakyRelu_217 (1, 32, 160, 160)
D Calc tensor Add_218 (1, 32, 160, 160)
D Calc tensor Conv_219 (1, 32, 160, 160)
D Calc tensor Conv_220 (1, 32, 160, 160)
D Calc tensor Concat_221 (1, 64, 160, 160)
D Calc tensor BatchNormalization_222 (1, 64, 160, 160)
D Calc tensor LeakyRelu_223 (1, 64, 160, 160)
D Calc tensor Conv_224 (1, 64, 160, 160)
D Calc tensor LeakyRelu_225 (1, 64, 160, 160)
D Calc tensor Conv_226 (1, 128, 80, 80)
D Calc tensor LeakyRelu_227 (1, 128, 80, 80)
D Calc tensor Conv_228 (1, 64, 80, 80)
D Calc tensor LeakyRelu_229 (1, 64, 80, 80)
D Calc tensor Conv_230 (1, 64, 80, 80)
D Calc tensor LeakyRelu_231 (1, 64, 80, 80)
D Calc tensor Conv_232 (1, 64, 80, 80)
D Calc tensor LeakyRelu_233 (1, 64, 80, 80)
D Calc tensor Add_234 (1, 64, 80, 80)
D Calc tensor Conv_235 (1, 64, 80, 80)
D Calc tensor LeakyRelu_236 (1, 64, 80, 80)
D Calc tensor Conv_237 (1, 64, 80, 80)
D Calc tensor LeakyRelu_238 (1, 64, 80, 80)
D Calc tensor Add_239 (1, 64, 80, 80)
D Calc tensor Conv_240 (1, 64, 80, 80)
D Calc tensor LeakyRelu_241 (1, 64, 80, 80)
D Calc tensor Conv_242 (1, 64, 80, 80)
D Calc tensor LeakyRelu_243 (1, 64, 80, 80)
D Calc tensor Add_244 (1, 64, 80, 80)
D Calc tensor Conv_245 (1, 64, 80, 80)
D Calc tensor Conv_246 (1, 64, 80, 80)
D Calc tensor Concat_247 (1, 128, 80, 80)
D Calc tensor BatchNormalization_248 (1, 128, 80, 80)
D Calc tensor LeakyRelu_249 (1, 128, 80, 80)
D Calc tensor Conv_250 (1, 128, 80, 80)
D Calc tensor LeakyRelu_251 (1, 128, 80, 80)
D Calc tensor Conv_252 (1, 256, 40, 40)
D Calc tensor LeakyRelu_253 (1, 256, 40, 40)
D Calc tensor Conv_254 (1, 128, 40, 40)
D Calc tensor LeakyRelu_255 (1, 128, 40, 40)
D Calc tensor Conv_256 (1, 128, 40, 40)
D Calc tensor LeakyRelu_257 (1, 128, 40, 40)
D Calc tensor Conv_258 (1, 128, 40, 40)
D Calc tensor LeakyRelu_259 (1, 128, 40, 40)
D Calc tensor Add_260 (1, 128, 40, 40)
D Calc tensor Conv_261 (1, 128, 40, 40)
D Calc tensor LeakyRelu_262 (1, 128, 40, 40)
D Calc tensor Conv_263 (1, 128, 40, 40)
D Calc tensor LeakyRelu_264 (1, 128, 40, 40)
D Calc tensor Add_265 (1, 128, 40, 40)
D Calc tensor Conv_266 (1, 128, 40, 40)
D Calc tensor LeakyRelu_267 (1, 128, 40, 40)
D Calc tensor Conv_268 (1, 128, 40, 40)
D Calc tensor LeakyRelu_269 (1, 128, 40, 40)
D Calc tensor Add_270 (1, 128, 40, 40)
D Calc tensor Conv_271 (1, 128, 40, 40)
D Calc tensor Conv_272 (1, 128, 40, 40)
D Calc tensor Concat_273 (1, 256, 40, 40)
D Calc tensor BatchNormalization_274 (1, 256, 40, 40)
D Calc tensor LeakyRelu_275 (1, 256, 40, 40)
D Calc tensor Conv_276 (1, 256, 40, 40)
D Calc tensor LeakyRelu_277 (1, 256, 40, 40)
D Calc tensor Conv_278 (1, 512, 20, 20)
D Calc tensor LeakyRelu_279 (1, 512, 20, 20)
D Calc tensor Conv_280 (1, 256, 20, 20)
D Calc tensor LeakyRelu_281 (1, 256, 20, 20)
D Calc tensor MaxPool_282 (1, 256, 20, 20)
D Calc tensor MaxPool_283 (1, 256, 20, 20)
D Calc tensor MaxPool_284 (1, 256, 20, 20)
D Calc tensor Concat_285 (1, 1024, 20, 20)
D Calc tensor Conv_286 (1, 512, 20, 20)
D Calc tensor LeakyRelu_287 (1, 512, 20, 20)
D Calc tensor Conv_288 (1, 256, 20, 20)
D Calc tensor LeakyRelu_289 (1, 256, 20, 20)
D Calc tensor Conv_290 (1, 256, 20, 20)
D Calc tensor LeakyRelu_291 (1, 256, 20, 20)
D Calc tensor Conv_292 (1, 256, 20, 20)
D Calc tensor LeakyRelu_293 (1, 256, 20, 20)
D Calc tensor Conv_294 (1, 256, 20, 20)
D Calc tensor Conv_295 (1, 256, 20, 20)
D Calc tensor Concat_296 (1, 512, 20, 20)
D Calc tensor BatchNormalization_297 (1, 512, 20, 20)
D Calc tensor LeakyRelu_298 (1, 512, 20, 20)
D Calc tensor Conv_299 (1, 512, 20, 20)
D Calc tensor LeakyRelu_300 (1, 512, 20, 20)
D Calc tensor Conv_301 (1, 256, 20, 20)
D Calc tensor LeakyRelu_302 (1, 256, 20, 20)
D Calc tensor Resize_331 (1, 256, 40, 40)
D Calc tensor Concat_332 (1, 512, 40, 40)
D Calc tensor Conv_333 (1, 128, 40, 40)
D Calc tensor LeakyRelu_334 (1, 128, 40, 40)
D Calc tensor Conv_335 (1, 128, 40, 40)
D Calc tensor LeakyRelu_336 (1, 128, 40, 40)
D Calc tensor Conv_337 (1, 128, 40, 40)
D Calc tensor LeakyRelu_338 (1, 128, 40, 40)
D Calc tensor Conv_339 (1, 128, 40, 40)
D Calc tensor Conv_340 (1, 128, 40, 40)
D Calc tensor Concat_341 (1, 256, 40, 40)
D Calc tensor BatchNormalization_342 (1, 256, 40, 40)
D Calc tensor LeakyRelu_343 (1, 256, 40, 40)
D Calc tensor Conv_344 (1, 256, 40, 40)
D Calc tensor LeakyRelu_345 (1, 256, 40, 40)
D Calc tensor Conv_346 (1, 128, 40, 40)
D Calc tensor LeakyRelu_347 (1, 128, 40, 40)
D Calc tensor Resize_376 (1, 128, 80, 80)
D Calc tensor Concat_377 (1, 256, 80, 80)
D Calc tensor Conv_378 (1, 64, 80, 80)
D Calc tensor LeakyRelu_379 (1, 64, 80, 80)
D Calc tensor Conv_380 (1, 64, 80, 80)
D Calc tensor LeakyRelu_381 (1, 64, 80, 80)
D Calc tensor Conv_382 (1, 64, 80, 80)
D Calc tensor LeakyRelu_383 (1, 64, 80, 80)
D Calc tensor Conv_384 (1, 64, 80, 80)
D Calc tensor Conv_385 (1, 64, 80, 80)
D Calc tensor Concat_386 (1, 128, 80, 80)
D Calc tensor BatchNormalization_387 (1, 128, 80, 80)
D Calc tensor LeakyRelu_388 (1, 128, 80, 80)
D Calc tensor Conv_389 (1, 128, 80, 80)
D Calc tensor LeakyRelu_390 (1, 128, 80, 80)
D Calc tensor Conv_391 (1, 128, 40, 40)
D Calc tensor LeakyRelu_392 (1, 128, 40, 40)
D Calc tensor Concat_393 (1, 256, 40, 40)
D Calc tensor Conv_394 (1, 128, 40, 40)
D Calc tensor LeakyRelu_395 (1, 128, 40, 40)
D Calc tensor Conv_396 (1, 128, 40, 40)
D Calc tensor LeakyRelu_397 (1, 128, 40, 40)
D Calc tensor Conv_398 (1, 128, 40, 40)
D Calc tensor LeakyRelu_399 (1, 128, 40, 40)
D Calc tensor Conv_400 (1, 128, 40, 40)
D Calc tensor Conv_401 (1, 128, 40, 40)
D Calc tensor Concat_402 (1, 256, 40, 40)
D Calc tensor BatchNormalization_403 (1, 256, 40, 40)
D Calc tensor LeakyRelu_404 (1, 256, 40, 40)
D Calc tensor Conv_405 (1, 256, 40, 40)
D Calc tensor LeakyRelu_406 (1, 256, 40, 40)
D Calc tensor Conv_407 (1, 256, 20, 20)
D Calc tensor LeakyRelu_408 (1, 256, 20, 20)
D Calc tensor Concat_409 (1, 512, 20, 20)
D Calc tensor Conv_410 (1, 256, 20, 20)
D Calc tensor LeakyRelu_411 (1, 256, 20, 20)
D Calc tensor Conv_412 (1, 256, 20, 20)
D Calc tensor LeakyRelu_413 (1, 256, 20, 20)
D Calc tensor Conv_414 (1, 256, 20, 20)
D Calc tensor LeakyRelu_415 (1, 256, 20, 20)
D Calc tensor Conv_416 (1, 256, 20, 20)
D Calc tensor Conv_417 (1, 256, 20, 20)
D Calc tensor Concat_418 (1, 512, 20, 20)
D Calc tensor BatchNormalization_419 (1, 512, 20, 20)
D Calc tensor LeakyRelu_420 (1, 512, 20, 20)
D Calc tensor Conv_421 (1, 512, 20, 20)
D Calc tensor LeakyRelu_422 (1, 512, 20, 20)
D Calc tensor Conv_463 (1, 24, 20, 20)
D Calc tensor Reshape_481 (1, 3, 8, 20, 20)
D Calc tensor Transpose_482 (1, 3, 20, 20, 8)
D Calc tensor Conv_443 (1, 24, 40, 40)
D Calc tensor Reshape_461 (1, 3, 8, 40, 40)
D Calc tensor Transpose_462 (1, 3, 40, 40, 8)
D Calc tensor Conv_423 (1, 24, 80, 80)
D Calc tensor Reshape_441 (1, 3, 8, 80, 80)
D Calc tensor Transpose_output (1, 3, 80, 80, 8)
D import clients finished
I build output layer attach_Transpose_Transpose_271ut0
I build output layer attach_Transpose_Transpose_287ut0
I build output layer attach_Transpose_Transpose_303ut0
I Try match Transpose_Transpose_303ut0
I Match r_transpose [['Transpose_Transpose_303']] [['Transpose']] to [['permute']]
I Try match Transpose_Transpose_287ut0
I Match r_transpose [['Transpose_Transpose_287']] [['Transpose']] to [['permute']]
I Try match Transpose_Transpose_271ut0
I Match r_transpose [['Transpose_Transpose_271']] [['Transpose']] to [['permute']]
I Try match Reshape_Reshape_302ut0
I Match r_rsp_v5 [['Reshape_Reshape_302', 'Initializer_480']] [['Reshape', 'Constant_0']] to [['reshape']]
I Try match Reshape_Reshape_286ut0
I Match r_rsp_v5 [['Reshape_Reshape_286', 'Initializer_460']] [['Reshape', 'Constant_0']] to [['reshape']]
I Try match Reshape_Reshape_270ut0
I Match r_rsp_v5 [['Reshape_Reshape_270', 'Initializer_440']] [['Reshape', 'Constant_0']] to [['reshape']]
I Try match Conv_Conv_288ut0
I Match r_conv [['Conv_Conv_288', 'Initializer_model.24.m.2.weight', 'Initializer_model.24.m.2.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match Conv_Conv_272:out0
I Match r_conv [['Conv_Conv_272', 'Initializer_model.24.m.1.weight', 'Initializer_model.24.m.1.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match Conv_Conv_256:out0
I Match r_conv [['Conv_Conv_256', 'Initializer_model.24.m.0.weight', 'Initializer_model.24.m.0.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match LeakyRelu_LeakyRelu_255:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_255']] [['LeakyRelu']] to [['leakyrelu']]
I Try match LeakyRelu_LeakyRelu_239:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_239']] [['LeakyRelu']] to [['leakyrelu']]
I Try match LeakyRelu_LeakyRelu_223:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_223']] [['LeakyRelu']] to [['leakyrelu']]
I Try match Conv_Conv_254:out0
I Match r_conv [['Conv_Conv_254', 'Initializer_model.23.cv4.conv.weight', 'Initializer_model.23.cv4.conv.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match Conv_Conv_238:out0
I Match r_conv [['Conv_Conv_238', 'Initializer_model.20.cv4.conv.weight', 'Initializer_model.20.cv4.conv.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match Conv_Conv_222:out0
I Match r_conv [['Conv_Conv_222', 'Initializer_model.17.cv4.conv.weight', 'Initializer_model.17.cv4.conv.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match LeakyRelu_LeakyRelu_253:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_253']] [['LeakyRelu']] to [['leakyrelu']]
I Try match LeakyRelu_LeakyRelu_237:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_237']] [['LeakyRelu']] to [['leakyrelu']]
I Try match LeakyRelu_LeakyRelu_221:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_221']] [['LeakyRelu']] to [['leakyrelu']]
I Try match BatchNormalization_BatchNormalization_252:out0
I Match r_bn_v6 [['BatchNormalization_BatchNormalization_252', 'Initializer_model.23.bn.weight', 'Initializer_model.23.bn.bias', 'Initializer_model.23.bn.running_mean', 'Initializer_model.23.bn.running_var']] [['BatchNormalization', 'Constant_0', 'Constant_1', 'Constant_2', 'Constant_3']] to [['batchnormalize']]
I Try match BatchNormalization_BatchNormalization_236:out0
I Match r_bn_v6 [['BatchNormalization_BatchNormalization_236', 'Initializer_model.20.bn.weight', 'Initializer_model.20.bn.bias', 'Initializer_model.20.bn.running_mean', 'Initializer_model.20.bn.running_var']] [['BatchNormalization', 'Constant_0', 'Constant_1', 'Constant_2', 'Constant_3']] to [['batchnormalize']]
I Try match BatchNormalization_BatchNormalization_220:out0
I Match r_bn_v6 [['BatchNormalization_BatchNormalization_220', 'Initializer_model.17.bn.weight', 'Initializer_model.17.bn.bias', 'Initializer_model.17.bn.running_mean', 'Initializer_model.17.bn.running_var']] [['BatchNormalization', 'Constant_0', 'Constant_1', 'Constant_2', 'Constant_3']] to [['batchnormalize']]
I Try match Concat_Concat_251:out0
I Match concat_2 [['Concat_Concat_251']] [['Concat']] to [['concat']]
I Try match Concat_Concat_235:out0
I Match concat_2 [['Concat_Concat_235']] [['Concat']] to [['concat']]
I Try match Concat_Concat_219:out0
I Match concat_2 [['Concat_Concat_219']] [['Concat']] to [['concat']]
I Try match Conv_Conv_249:out0
I Match r_conv_no_bias [['Conv_Conv_249', 'Initializer_model.23.cv3.weight']] [['Conv', 'Constant_0']] to [['convolution']]
I Try match Conv_Conv_250:out0
I Match r_conv_no_bias [['Conv_Conv_250', 'Initializer_model.23.cv2.weight']] [['Conv', 'Constant_0']] to [['convolution']]
I Try match Conv_Conv_233:out0
I Match r_conv_no_bias [['Conv_Conv_233', 'Initializer_model.20.cv3.weight']] [['Conv', 'Constant_0']] to [['convolution']]
I Try match Conv_Conv_234:out0
I Match r_conv_no_bias [['Conv_Conv_234', 'Initializer_model.20.cv2.weight']] [['Conv', 'Constant_0']] to [['convolution']]
I Try match Conv_Conv_217:out0
I Match r_conv_no_bias [['Conv_Conv_217', 'Initializer_model.17.cv3.weight']] [['Conv', 'Constant_0']] to [['convolution']]
I Try match Conv_Conv_218:out0
I Match r_conv_no_bias [['Conv_Conv_218', 'Initializer_model.17.cv2.weight']] [['Conv', 'Constant_0']] to [['convolution']]
I Try match LeakyRelu_LeakyRelu_248:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_248']] [['LeakyRelu']] to [['leakyrelu']]
I Try match Concat_Concat_242:out0
I Match concat_2 [['Concat_Concat_242']] [['Concat']] to [['concat']]
I Try match LeakyRelu_LeakyRelu_232:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_232']] [['LeakyRelu']] to [['leakyrelu']]
I Try match Concat_Concat_226:out0
I Match concat_2 [['Concat_Concat_226']] [['Concat']] to [['concat']]
I Try match LeakyRelu_LeakyRelu_216:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_216']] [['LeakyRelu']] to [['leakyrelu']]
I Try match Concat_Concat_210:out0
I Match concat_2 [['Concat_Concat_210']] [['Concat']] to [['concat']]
I Try match Conv_Conv_247:out0
I Match r_conv [['Conv_Conv_247', 'Initializer_model.23.m.0.cv2.conv.weight', 'Initializer_model.23.m.0.cv2.conv.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match LeakyRelu_LeakyRelu_241:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_241']] [['LeakyRelu']] to [['leakyrelu']]
I Try match LeakyRelu_LeakyRelu_135:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_135']] [['LeakyRelu']] to [['leakyrelu']]
I Try match Conv_Conv_231:out0
I Match r_conv [['Conv_Conv_231', 'Initializer_model.20.m.0.cv2.conv.weight', 'Initializer_model.20.m.0.cv2.conv.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match LeakyRelu_LeakyRelu_225:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_225']] [['LeakyRelu']] to [['leakyrelu']]
I Try match LeakyRelu_LeakyRelu_180:out0
I Match r_leakrelu [['LeakyRelu_LeakyRelu_180']] [['LeakyRelu']] to [['leakyrelu']]
I Try match Conv_Conv_215:out0
I Match r_conv [['Conv_Conv_215', 'Initializer_model.17.m.0.cv2.conv.weight', 'Initializer_model.17.m.0.cv2.conv.bias']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]
I Try match Resize_Resize_209:out0
W Not match tensor Resize_Resize_209:out0
E Try match Resize_Resize_209:out0 failed, catch exception!
W ----------------Warning(3)----------------
E Catch exception when loading onnx model: ./weights/best_sim.onnx!
E Traceback (most recent call last):
E   File "rknn\base\RKNNlib\converter\convert_onnx.py", line 826, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx.match_paragraph_and_param
E   File "rknn\base\RKNNlib\converter\convert_onnx.py", line 728, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx._onnx_push_ready_tensor
E TypeError: 'NoneType' object is not iterable
E During handling of the above exception, another exception occurred:
E Traceback (most recent call last):
E   File "rknn\api\rknn_base.py", line 264, in rknn.api.rknn_base.RKNNBase.load_onnx
E   File "rknn\base\RKNNlib\RK_nn.py", line 135, in rknn.base.RKNNlib.RK_nn.RKnn.load_onnx
E   File "rknn\base\RKNNlib\app\importer\import_onnx.py", line 121, in rknn.base.RKNNlib.app.importer.import_onnx.Importonnx.run
E   File "rknn\base\RKNNlib\converter\convert_onnx.py", line 832, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx.match_paragraph_and_param
E   File "rknn\api\rknn_log.py", line 312, in rknn.api.rknn_log.RKNNLog.e
E ValueError: Try match Resize_Resize_209:out0 failed, catch exception!
Load model failed!

我的加载模型代码：
ONNX_MODEL = './weights/best_sim.onnx'
    RKNN_MODEL = 'best_640x640.rknn'

    # Create RKNN object
    rknn = RKNN(verbose=True)
    print('--> config model')
    rknn.config(channel_mean_value='0 0 0 255', reorder_channel='0 1 2', batch_size=1)
    print('done')

    # Load  model
    print('--> Loading model')
    model = ONNX_MODEL

    ret = rknn.load_onnx(model=model)
    if ret != 0:
        print('Load model failed!')
        exit(ret)
    print('done')

我的模型是YOLOv5模型，但我尝试直接用pt文件转换为rknn的时候，又出现了如下问题：（还是加载模型的错误）
W Verbose file path is invalid, debug info will not dump to file.
--> Config model
done
--> Loading model
I Start importing pytorch...
./weights/best.pt ********************
D import clients finished
E Catch exception when loading pytorch model: ./weights/best.pt!
E Traceback (most recent call last):
E   File "rknn\api\rknn_base.py", line 339, in rknn.api.rknn_base.RKNNBase.load_pytorch
E   File "rknn\base\RKNNlib\RK_nn.py", line 146, in rknn.base.RKNNlib.RK_nn.RKnn.load_pytorch
E   File "rknn\base\RKNNlib\app\importer\import_pytorch.py", line 128, in rknn.base.RKNNlib.app.importer.import_pytorch.ImportPytorch.run
E   File "rknn\base\RKNNlib\converter\convert_pytorch_new.py", line 2243, in rknn.base.RKNNlib.converter.convert_pytorch_new.convert_pytorch.load
E   File "D:\Anaconda3\envs\py363\lib\site-packages\torch\jit\__init__.py", line 228, in load
E     cpp_module = torch._C.import_ir_module(cu, f, map_location, _extra_files)
E RuntimeError: [enforce fail at ..\caffe2\serialize\inline_container.cc:143] . PytorchStreamReader failed reading zip archive: failed finding central directory
Load Pytorch model failed!
我的加载代码如下：
   weight = './weights/best.pt'

    input_size_list = [[3, 640, 640]]

    # Create RKNN object
    rknn = RKNN(verbose=True)

    # pre-process config
    print('--> Config model')
    # rknn.config(mean_values=[[123.675, 116.28, 103.53]], std_values=[[58.395, 58.395, 58.395]], reorder_channel='0 1 2')
    rknn.config(channel_mean_value='0 0 0 255', reorder_channel='0 1 2', batch_size=1)
    print('done')

    # Load Pytorch model
    print('--> Loading model')
    ret = rknn.load_pytorch(model=weight, input_size_list=input_size_list)
    if ret != 0:
        print('Load Pytorch model failed!')
        exit(ret)
    print('done')

yuelaiyue 发表于 2021-3-17 10:05
我的模型是YOLOv5模型，但我尝试直接用pt文件转换为rknn的时候，又出现了如下问题：（还是加载模型的错误） ...

解决了吗？，速度如何？

Try match Resize_Resize_209ut0 failed,
这个OP不支持

pytorchd导出onnx时候 设置opset_version=10