При подготовке шаблона для конвертера MMdnn перестало работать чтение модели
valentina-kustikova opened this issue · comments
При подготовке шаблона для конвертера MMdnn перестало работать чтение модели средствами конвертера, который лежал в model_converters
Командная строка:
python inference_tensorflow.py --model /home/itmm/Documents/kustikova_v/public/densenet-121-tf/densenet-121.savedmodel/ --input /home/itmm/Documents/kustikova_v/data/ --batch_size 3 --task classificationОшибка:
Traceback (most recent call last):
File "inference_tensorflow.py", line 19, in <module>
from src.model_converters.tf2tflite.tensorflow_common import (load_model, get_gpu_devices, is_gpu_available, # noqa
ModuleNotFoundError: No module named 'src'
@valentina-kustikova у меня версия из мастера с моделью densenet-121-tf работает, ошибки src не выдает. Добавь пожалуйста больше информации про текущий энвайрмент - версию Python и TensorFlow. И еще информацию про то, из какой директории вызывается скрипт.
У меня: Python 3.9.15, TensorFlow 2.11.0, вызываю из папки inference
python3 inference_tensorflow.py --model public/densenet-121-tf/densenet-121.savedmodel --input ../../results/validation/images/ILSVRC2012_val_00018592.JPEG -b 1 --input_shape 224 224 3 -ni 100
[ INFO ] Loading network files:
public/densenet-121-tf/densenet-121.savedmodel
[ INFO ] Loading saved model from public/densenet-121-tf/densenet-121.savedmodel
2023-11-06 22:28:42.538875: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
[ INFO ] Shape for input layer input_1:0: 1x224x224x3
<вырезал ворнинги>
[ INFO ] Preparing input data: ['../../results/validation/images/ILSVRC2012_val_00018592.JPEG']
[ INFO ] Got input names ['input_1:0', 'unknown:0', 'unknown_0:0', 'unknown_1:0', 'unknown_2:0', 'unknown_3:0', 'unknown_4:0', 'unknown_5:0', 'unknown_6:0', 'unknown_7:0', 'unknown_8:0', 'unknown_9:0', 'unknown_10:0', 'unknown_11:0', 'unknown_12:0', 'unknown_13:0', 'unknown_14:0', 'unknown_15:0', 'unknown_16:0', 'unknown_17:0', 'unknown_18:0', 'unknown_19:0', 'unknown_20:0', 'unknown_21:0', 'unknown_22:0', 'unknown_23:0', 'unknown_24:0', 'unknown_25:0', 'unknown_26:0', 'unknown_27:0', 'unknown_28:0', 'unknown_29:0', 'unknown_30:0', 'unknown_31:0', 'unknown_32:0', 'unknown_33:0', 'unknown_34:0', 'unknown_35:0', 'unknown_36:0', 'unknown_37:0', 'unknown_38:0', 'unknown_39:0', 'unknown_40:0', 'unknown_41:0', 'unknown_42:0', 'unknown_43:0', 'unknown_44:0', 'unknown_45:0', 'unknown_46:0', 'unknown_47:0', 'unknown_48:0', 'unknown_49:0', 'unknown_50:0', 'unknown_51:0', 'unknown_52:0', 'unknown_53:0', 'unknown_54:0', 'unknown_55:0', 'unknown_56:0', 'unknown_57:0', 'unknown_58:0', 'unknown_59:0', 'unknown_60:0', 'unknown_61:0', 'unknown_62:0', 'unknown_63:0', 'unknown_64:0', 'unknown_65:0', 'unknown_66:0', 'unknown_67:0', 'unknown_68:0', 'unknown_69:0', 'unknown_70:0', 'unknown_71:0', 'unknown_72:0', 'unknown_73:0', 'unknown_74:0', 'unknown_75:0', 'unknown_76:0', 'unknown_77:0', 'unknown_78:0', 'unknown_79:0', 'unknown_80:0', 'unknown_81:0', 'unknown_82:0', 'unknown_83:0', 'unknown_84:0', 'unknown_85:0', 'unknown_86:0', 'unknown_87:0', 'unknown_88:0', 'unknown_89:0', 'unknown_90:0', 'unknown_91:0', 'unknown_92:0', 'unknown_93:0', 'unknown_94:0', 'unknown_95:0', 'unknown_96:0', 'unknown_97:0', 'unknown_98:0', 'unknown_99:0', 'unknown_100:0', 'unknown_101:0', 'unknown_102:0', 'unknown_103:0', 'unknown_104:0', 'unknown_105:0', 'unknown_106:0', 'unknown_107:0', 'unknown_108:0', 'unknown_109:0', 'unknown_110:0', 'unknown_111:0', 'unknown_112:0', 'unknown_113:0', 'unknown_114:0', 'unknown_115:0', 'unknown_116:0', 'unknown_117:0', 'unknown_118:0', 'unknown_119:0', 'unknown_120:0', 'unknown_121:0', 'unknown_122:0', 'unknown_123:0', 'unknown_124:0', 'unknown_125:0', 'unknown_126:0', 'unknown_127:0', 'unknown_128:0', 'unknown_129:0', 'unknown_130:0', 'unknown_131:0', 'unknown_132:0', 'unknown_133:0', 'unknown_134:0', 'unknown_135:0', 'unknown_136:0', 'unknown_137:0', 'unknown_138:0', 'unknown_139:0', 'unknown_140:0', 'unknown_141:0', 'unknown_142:0', 'unknown_143:0', 'unknown_144:0', 'unknown_145:0', 'unknown_146:0', 'unknown_147:0', 'unknown_148:0', 'unknown_149:0', 'unknown_150:0', 'unknown_151:0', 'unknown_152:0', 'unknown_153:0', 'unknown_154:0', 'unknown_155:0', 'unknown_156:0', 'unknown_157:0', 'unknown_158:0', 'unknown_159:0', 'unknown_160:0', 'unknown_161:0', 'unknown_162:0', 'unknown_163:0', 'unknown_164:0', 'unknown_165:0', 'unknown_166:0', 'unknown_167:0', 'unknown_168:0', 'unknown_169:0', 'unknown_170:0', 'unknown_171:0', 'unknown_172:0', 'unknown_173:0', 'unknown_174:0', 'unknown_175:0', 'unknown_176:0', 'unknown_177:0', 'unknown_178:0', 'unknown_179:0', 'unknown_180:0', 'unknown_181:0', 'unknown_182:0', 'unknown_183:0', 'unknown_184:0', 'unknown_185:0', 'unknown_186:0', 'unknown_187:0', 'unknown_188:0', 'unknown_189:0', 'unknown_190:0', 'unknown_191:0', 'unknown_192:0', 'unknown_193:0', 'unknown_194:0', 'unknown_195:0', 'unknown_196:0', 'unknown_197:0', 'unknown_198:0', 'unknown_199:0', 'unknown_200:0', 'unknown_201:0', 'unknown_202:0', 'unknown_203:0', 'unknown_204:0', 'unknown_205:0', 'unknown_206:0', 'unknown_207:0', 'unknown_208:0', 'unknown_209:0', 'unknown_210:0', 'unknown_211:0', 'unknown_212:0', 'unknown_213:0', 'unknown_214:0', 'unknown_215:0', 'unknown_216:0', 'unknown_217:0', 'unknown_218:0', 'unknown_219:0', 'unknown_220:0', 'unknown_221:0', 'unknown_222:0', 'unknown_223:0', 'unknown_224:0', 'unknown_225:0', 'unknown_226:0', 'unknown_227:0', 'unknown_228:0', 'unknown_229:0', 'unknown_230:0', 'unknown_231:0', 'unknown_232:0', 'unknown_233:0', 'unknown_234:0', 'unknown_235:0', 'unknown_236:0', 'unknown_237:0', 'unknown_238:0', 'unknown_239:0', 'unknown_240:0', 'unknown_241:0', 'unknown_242:0', 'unknown_243:0', 'unknown_244:0', 'unknown_245:0', 'unknown_246:0', 'unknown_247:0', 'unknown_248:0', 'unknown_249:0', 'unknown_250:0', 'unknown_251:0', 'unknown_252:0', 'unknown_253:0', 'unknown_254:0', 'unknown_255:0', 'unknown_256:0', 'unknown_257:0', 'unknown_258:0', 'unknown_259:0', 'unknown_260:0', 'unknown_261:0', 'unknown_262:0', 'unknown_263:0', 'unknown_264:0', 'unknown_265:0', 'unknown_266:0', 'unknown_267:0', 'unknown_268:0', 'unknown_269:0', 'unknown_270:0', 'unknown_271:0', 'unknown_272:0', 'unknown_273:0', 'unknown_274:0', 'unknown_275:0', 'unknown_276:0', 'unknown_277:0', 'unknown_278:0', 'unknown_279:0', 'unknown_280:0', 'unknown_281:0', 'unknown_282:0', 'unknown_283:0', 'unknown_284:0', 'unknown_285:0', 'unknown_286:0', 'unknown_287:0', 'unknown_288:0', 'unknown_289:0', 'unknown_290:0', 'unknown_291:0', 'unknown_292:0', 'unknown_293:0', 'unknown_294:0', 'unknown_295:0', 'unknown_296:0', 'unknown_297:0', 'unknown_298:0', 'unknown_299:0', 'unknown_300:0', 'unknown_301:0', 'unknown_302:0', 'unknown_303:0', 'unknown_304:0', 'unknown_305:0', 'unknown_306:0', 'unknown_307:0', 'unknown_308:0', 'unknown_309:0', 'unknown_310:0', 'unknown_311:0', 'unknown_312:0', 'unknown_313:0', 'unknown_314:0', 'unknown_315:0', 'unknown_316:0', 'unknown_317:0', 'unknown_318:0', 'unknown_319:0', 'unknown_320:0', 'unknown_321:0', 'unknown_322:0', 'unknown_323:0', 'unknown_324:0', 'unknown_325:0', 'unknown_326:0', 'unknown_327:0', 'unknown_328:0', 'unknown_329:0', 'unknown_330:0', 'unknown_331:0', 'unknown_332:0', 'unknown_333:0', 'unknown_334:0', 'unknown_335:0', 'unknown_336:0', 'unknown_337:0', 'unknown_338:0', 'unknown_339:0', 'unknown_340:0', 'unknown_341:0', 'unknown_342:0', 'unknown_343:0', 'unknown_344:0', 'unknown_345:0', 'unknown_346:0', 'unknown_347:0', 'unknown_348:0', 'unknown_349:0', 'unknown_350:0', 'unknown_351:0', 'unknown_352:0', 'unknown_353:0', 'unknown_354:0', 'unknown_355:0', 'unknown_356:0', 'unknown_357:0', 'unknown_358:0', 'unknown_359:0', 'unknown_360:0', 'unknown_361:0', 'unknown_362:0', 'unknown_363:0', 'unknown_364:0', 'unknown_365:0', 'unknown_366:0', 'unknown_367:0', 'unknown_368:0', 'unknown_369:0', 'unknown_370:0', 'unknown_371:0', 'unknown_372:0', 'unknown_373:0', 'unknown_374:0', 'unknown_375:0', 'unknown_376:0', 'unknown_377:0', 'unknown_378:0', 'unknown_379:0', 'unknown_380:0', 'unknown_381:0', 'unknown_382:0', 'unknown_383:0', 'unknown_384:0', 'unknown_385:0', 'unknown_386:0', 'unknown_387:0', 'unknown_388:0', 'unknown_389:0', 'unknown_390:0', 'unknown_391:0', 'unknown_392:0', 'unknown_393:0', 'unknown_394:0', 'unknown_395:0', 'unknown_396:0', 'unknown_397:0', 'unknown_398:0', 'unknown_399:0', 'unknown_400:0', 'unknown_401:0', 'unknown_402:0', 'unknown_403:0', 'unknown_404:0', 'unknown_405:0', 'unknown_406:0', 'unknown_407:0', 'unknown_408:0', 'unknown_409:0', 'unknown_410:0', 'unknown_411:0', 'unknown_412:0', 'unknown_413:0', 'unknown_414:0', 'unknown_415:0', 'unknown_416:0', 'unknown_417:0', 'unknown_418:0', 'unknown_419:0', 'unknown_420:0', 'unknown_421:0', 'unknown_422:0', 'unknown_423:0', 'unknown_424:0', 'unknown_425:0', 'unknown_426:0', 'unknown_427:0', 'unknown_428:0', 'unknown_429:0', 'unknown_430:0', 'unknown_431:0', 'unknown_432:0', 'unknown_433:0', 'unknown_434:0', 'unknown_435:0', 'unknown_436:0', 'unknown_437:0', 'unknown_438:0', 'unknown_439:0', 'unknown_440:0', 'unknown_441:0', 'unknown_442:0', 'unknown_443:0', 'unknown_444:0', 'unknown_445:0', 'unknown_446:0', 'unknown_447:0', 'unknown_448:0', 'unknown_449:0', 'unknown_450:0', 'unknown_451:0', 'unknown_452:0', 'unknown_453:0', 'unknown_454:0', 'unknown_455:0', 'unknown_456:0', 'unknown_457:0', 'unknown_458:0', 'unknown_459:0', 'unknown_460:0', 'unknown_461:0', 'unknown_462:0', 'unknown_463:0', 'unknown_464:0', 'unknown_465:0', 'unknown_466:0', 'unknown_467:0', 'unknown_468:0', 'unknown_469:0', 'unknown_470:0', 'unknown_471:0', 'unknown_472:0', 'unknown_473:0', 'unknown_474:0', 'unknown_475:0', 'unknown_476:0', 'unknown_477:0', 'unknown_478:0', 'unknown_479:0', 'unknown_480:0', 'unknown_481:0', 'unknown_482:0', 'unknown_483:0', 'unknown_484:0', 'unknown_485:0', 'unknown_486:0', 'unknown_487:0', 'unknown_488:0', 'unknown_489:0', 'unknown_490:0', 'unknown_491:0', 'unknown_492:0', 'unknown_493:0', 'unknown_494:0', 'unknown_495:0', 'unknown_496:0', 'unknown_497:0', 'unknown_498:0', 'unknown_499:0', 'unknown_500:0', 'unknown_501:0', 'unknown_502:0', 'unknown_503:0', 'unknown_504:0', 'unknown_505:0', 'unknown_506:0', 'unknown_507:0', 'unknown_508:0', 'unknown_509:0', 'unknown_510:0', 'unknown_511:0', 'unknown_512:0', 'unknown_513:0', 'unknown_514:0', 'unknown_515:0', 'unknown_516:0', 'unknown_517:0', 'unknown_518:0', 'unknown_519:0', 'unknown_520:0', 'unknown_521:0', 'unknown_522:0', 'unknown_523:0', 'unknown_524:0', 'unknown_525:0', 'unknown_526:0', 'unknown_527:0', 'unknown_528:0', 'unknown_529:0', 'unknown_530:0', 'unknown_531:0', 'unknown_532:0', 'unknown_533:0', 'unknown_534:0', 'unknown_535:0', 'unknown_536:0', 'unknown_537:0', 'unknown_538:0', 'unknown_539:0', 'unknown_540:0', 'unknown_541:0', 'unknown_542:0', 'unknown_543:0', 'unknown_544:0', 'unknown_545:0', 'unknown_546:0', 'unknown_547:0', 'unknown_548:0', 'unknown_549:0', 'unknown_550:0', 'unknown_551:0', 'unknown_552:0', 'unknown_553:0', 'unknown_554:0', 'unknown_555:0', 'unknown_556:0', 'unknown_557:0', 'unknown_558:0', 'unknown_559:0', 'unknown_560:0', 'unknown_561:0', 'unknown_562:0', 'unknown_563:0', 'unknown_564:0', 'unknown_565:0', 'unknown_566:0', 'unknown_567:0', 'unknown_568:0', 'unknown_569:0', 'unknown_570:0', 'unknown_571:0', 'unknown_572:0', 'unknown_573:0', 'unknown_574:0', 'unknown_575:0', 'unknown_576:0', 'unknown_577:0', 'unknown_578:0', 'unknown_579:0', 'unknown_580:0', 'unknown_581:0', 'unknown_582:0', 'unknown_583:0', 'unknown_584:0', 'unknown_585:0', 'unknown_586:0', 'unknown_587:0', 'unknown_588:0', 'unknown_589:0', 'unknown_590:0', 'unknown_591:0', 'unknown_592:0', 'unknown_593:0', 'unknown_594:0', 'unknown_595:0', 'unknown_596:0', 'unknown_597:0', 'unknown_598:0', 'unknown_599:0', 'unknown_600:0', 'unknown_601:0', 'unknown_602:0', 'unknown_603:0', 'unknown_604:0']
[ INFO ] Starting inference (100 iterations)
start inference max 100 iterations or 0 seconds
....................................................................................................
[ INFO ] Inference completed
[ INFO ] Computing performance metrics
[ INFO ] Write report to /home/vasiliev/_dev/itlab/dl-benchmark/src/inference/tf_inference_report.json
[ INFO ] Performance results:
{
"iterations_num": 100,
"execution_time": 2.847,
"first_inference_time": 0.37946,
"latency_avg": 0.02492,
"latency_median": 0.02475,
"latency_std": 0.00086,
"latency_max": 0.02741,
"latency_min": 0.02327,
"batch_throughput": 40.401,
"throughput": 40.528
}
Версии моих пакетов:
tensorflow 2.9.3
Python 3.8.0
Вызываю вывод также из директории inference.
Странно как-то, но поставила python 3.9 и tensorflow 2.14.0, и проблема ушла.