Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

running on cpu error #6

Open
nparkstar opened this issue Mar 8, 2020 · 0 comments
Open

running on cpu error #6

nparkstar opened this issue Mar 8, 2020 · 0 comments

Comments

@nparkstar
Copy link

nparkstar commented Mar 8, 2020

트레이닝과 인퍼런스를 cpu에서 실행하려 하는데 에러가 납니다.
아래와 같은 메시지가 나오는데 검색을 해보니 Tensorflow가 gather_nd 메소드를 실행할 때 GPU에서는 경계 조건 위반 에러를 처리하지 않고 무조건 0으로 바꾸어서 처리하는데 반해서 CPU에서는 에러처리를 한다고 합니다. 이 것과 관련된 문제로 추정이되는데 어떤 값을 수정해야 문제가 해결리 되는지 알수가 없어서 문의드립니다.
CPU상에서 트레이닝이나 인퍼런스를 하면 모두 비슷한 문제가 발생합니다. 문제를 재현하는 가장 간단한 방법은
rfaster_rcnn_meta_arch_test_lib.py를 실행하면 249행
tensor_dict_out = sess.run(result_tensor_dict, feed_dict={preprocessed_inputs: np.zeros((batch_size, image_size, image_size, 3))})
가 실행될 때 에러가 납니다.
------------ 에러 메시지 -------------------
2020-03-06 13:23:55.105126: W tensorflow/core/framework/op_kernel.cc:1651] OP_REQUIRES failed at gather_nd_op.cc:47 : Invalid argument: indices[22,15,0] = [22, 15, -1] does not index into param shape [24,24,24]
2020-03-06 13:23:55.105158: W tensorflow/core/framework/op_kernel.cc:1651] OP_REQUIRES failed at gather_nd_op.cc:47 : Invalid argument: indices[22,15,0] = [22, 15, -1] does not index into param shape [24,24,24]
2020-03-06 13:23:55.105511: W tensorflow/core/framework/op_kernel.cc:1651] OP_REQUIRES failed at gather_nd_op.cc:47 : Invalid argument: indices[22,15,0] = [22, 15, -1] does not index into param shape [24,24,24]
2020-03-06 13:23:55.105539: W tensorflow/core/framework/op_kernel.cc:1651] OP_REQUIRES failed at gather_nd_op.cc:47 : Invalid argument: indices[22,15,0] = [22, 15, -1] does not index into param shape [24,24,24]
ERROR:tensorflow:indices[22,15,0] = [22, 15, -1] does not index into param shape [24,24,24]
[[node BatchMultiClassNonMaxSuppression/map/while/MultiClassNonMaxSuppression/NonMaxSuppressionRbox/IOURbox/cond/IntersectionRboxTF_1/cond/GatherNd_3 (defined at /root/anaconda3/envs/intel-tf/lib/python3.6/site-packages/tensorflow_core/python/framework/ops.py:1748) ]]

Original stack trace for 'BatchMultiClassNonMaxSuppression/map/while/MultiClassNonMaxSuppression/NonMaxSuppressionRbox/IOURbox/cond/IntersectionRboxTF_1/cond/GatherNd_3':

감사합니다.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant