Я использую tenorflow-gpu (1.12.0), python3 .6 (anaconda) для обнаружения объектов на windows 10.
Когда я использую эту команду для обучения моей модели:
python model_main.py --pipeline_config_path = training / ssd_inception_v2_coco.config --model_dir = training / --num_train_steps = 200000 --alsologtostderr
Произошла ошибка. Сообщение об ошибке:
2020-01-09 10:45:21.136763: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
2020-01-09 10:45:21.626182: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1432] Found device 0 with properties:
name: GeForce GTX 1050 major: 6 minor: 1 memoryClockRate(GHz): 1.493
pciBusID: 0000:01:00.0
totalMemory: 4.00GiB freeMemory: 3.30GiB
2020-01-09 10:45:21.631898: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1511] Adding visible gpu devices: 0
2020-01-09 10:45:22.525168: I tensorflow/core/common_runtime/gpu/gpu_device.cc:982] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-01-09 10:45:22.529085: I tensorflow/core/common_runtime/gpu/gpu_device.cc:988] 0
2020-01-09 10:45:22.530525: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1001] 0: N
2020-01-09 10:45:22.532224: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 3011 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1050, pci bus id: 0000:01
:00.0, compute capability: 6.1)
2020-01-09 10:45:55.213071: E tensorflow/stream_executor/cuda/cuda_dnn.cc:363] Loaded runtime CuDNN library: 7.0.5 but source was compiled with: 7.2.1. CuDNN library major and minor version needs to match or have higher minor version in case of CuD
NN 7.0 or later version. If using a binary install, upgrade your CuDNN library. If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.
2020-01-09 10:45:55.221967: E tensorflow/stream_executor/cuda/cuda_dnn.cc:363] Loaded runtime CuDNN library: 7.0.5 but source was compiled with: 7.2.1. CuDNN library major and minor version needs to match or have higher minor version in case of CuD
NN 7.0 or later version. If using a binary install, upgrade your CuDNN library. If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
model_main.py 109 <module>
tf.app.run()
app.py 125 run
_sys.exit(main(argv))
model_main.py 105 main
tf.estimator.train_and_evaluate(estimator, train_spec, eval_specs[0])
training.py 471 train_and_evaluate
return executor.run()
training.py 610 run
return self.run_local()
training.py 711 run_local
saving_listeners=saving_listeners)
estimator.py 354 train
loss = self._train_model(input_fn, hooks, saving_listeners)
estimator.py 1207 _train_model
return self._train_model_default(input_fn, hooks, saving_listeners)
estimator.py 1241 _train_model_default
saving_listeners)
estimator.py 1471 _train_with_estimator_spec
_, loss = mon_sess.run([estimator_spec.train_op, estimator_spec.loss])
monitored_session.py 671 run
run_metadata=run_metadata)
monitored_session.py 1156 run
run_metadata=run_metadata)
monitored_session.py 1255 run
raise six.reraise(*original_exc_info)
six.py 696 reraise
raise value
monitored_session.py 1240 run
return self._sess.run(*args, **kwargs)
monitored_session.py 1312 run
run_metadata=run_metadata)
monitored_session.py 1076 run
return self._sess.run(*args, **kwargs)
session.py 929 run
run_metadata_ptr)
session.py 1152 _run
feed_dict_tensor, options, run_metadata)
session.py 1328 _do_run
run_metadata)
session.py 1348 _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.UnknownError
Как исправить эту ошибку?