【发布时间】：2020-03-04 06:13:43
【问题描述】：
无论我使用object_detection/legacy/eval.py 脚本还是object_detection/model_main.py，每当评估开始时，我都会遇到分段错误。
我尝试重新安装 tensorflow、protobuf-compiler 并重新安装 tensorflow/modules 对象检测 api 的所有依赖项，但没有帮助。我正在使用 python 3。
Tensorflow 版本：1.15.0
我调用 eval.py 的方式：
python object_detection/legacy/eval.py \
--logtostderr \
--eval_dir=${TRAIN_DIR} \
--checkpoint_dir=${TRAIN_DIR} \
--pipeline_config_path=${PIPELINE_CONFIG_PATH}
这是运行脚本时的日志段
I1107 17:50:24.818984 140633596692288 saver.py:1284] Restoring parameters from /home/ihahanov/Projects/rooftops/resources/models/mask_rcnn_inception_v2_coco_2018_01_28/train/model.ckpt-29722
2019-11-07 17:50:27.309148: W tensorflow/core/framework/cpu_allocator_impl.cc:81] Allocation of 92160000 exceeds 10% of system memory.
Fatal Python error: Segmentation fault
Thread 0x00007fe6c4ff9700 (most recent call first):
File "/usr/lib/python3.6/threading.py", line 295 in wait
File "/usr/lib/python3.6/threading.py", line 551 in wait
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/training/coordinator.py", line 311 in wait_for_stop
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/training/queue_runner_impl.py", line 293 in _close_on_stop
File "/usr/lib/python3.6/threading.py", line 864 in run
File "/usr/lib/python3.6/threading.py", line 916 in _bootstrap_inner
File "/usr/lib/python3.6/threading.py", line 884 in _bootstrap
Thread 0x00007fe6bbfff700 (most recent call first):
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1443 in _call_tf_sessionrun
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1287 in _single_operation_run
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/training/queue_runner_impl.py", line 257 in _run
File "/usr/lib/python3.6/threading.py", line 864 in run
File "/usr/lib/python3.6/threading.py", line 916 in _bootstrap_inner
File "/usr/lib/python3.6/threading.py", line 884 in _bootstrap
Thread 0x00007fe7cf930740 (most recent call first):
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1443 in _call_tf_sessionrun
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1350 in _run_fn
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1365 in _do_call
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1359 in _do_run
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1180 in _run
File "/home/ihahanov/Projects/rooftops/venv/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 956 in run
File "/home/ihahanov/Projects/rooftops/models/research/object_detection/legacy/evaluator.py", line 234 in _process_batch
File "/home/ihahanov/Projects/rooftops/models/research/object_detection/eval_util.py", line 346 in _run_checkpoint_once
File "/home/ihahanov/Projects/rooftops/models/research/o/home/ihahanov/Projects/rooftops/shell/train_model.sh: line 26: 13874 Segmentation fault (core dumped) python3 eval.py --logtostderr --eval_dir=${TRAIN_DIR} --checkpoint_dir=${TRAIN_DIR} --pipeline_config_path=${PIPELINE_CONFIG_PATH}
【问题讨论】：
-
即使我面临同样的问题，我已经用 CPU 替换了 GPU 版本的 Tensorflow，并且没有出现任何错误。我检查了 CUDA（需要 10.0 但有 10.1）和 Nvidia 驱动程序（需要 418 和有 430）在我的情况下这些看起来不兼容。
-
实际上，当我开始在 GPU 而不是 CPU 上进行评估时，这个错误就消失了。当它是 CPU 时，错误发生在 3 台不同的机器上。
标签： python python-3.x tensorflow object-detection tensorflow-serving