formath/tensorflow-predictor-cpp

Not found: Op type not registered 'HashTableV2'

mmmwhy opened this issue · 5 comments

mmmwhy commented

sh predict.sh

Session created successfully
Load graph protobuf successfully
Not found: Op type not registered 'HashTableV2' in binary running on MacBook-Pro.local. Make sure the Op and Kernel are registered in the binary running in this process. Note that if you are loading a saved graph which used ops from tf.contrib, accessing (e.g.) tf.contrib.resampler should be done before importing the graph, as contrib ops are lazily registered when the module is first accessed.

环境：
tensorflow 1.8
python 3.6

mmmwhy commented

已尝试使用 https://github.com/formath/tensorflow.git branch: fix 编译，还是报错。

formath commented

What is the error using branch: fix?

mmmwhy commented

i have rewritten this inference part, thank you.

mjjdick commented

@mmmwhy @formath I met the same error!
environment:
tensorflow 1.10
python 2.7
it seems the bug about sparse tensor? can you guys help me? thx very much!

xixiddd commented

the same problem using "fix" branch

Session created successfully
Load graph protobuf successfully
Not found: Op type not registered 'HashTableV2' in binary running on a00bbeef0667. Make sure the Op and Kernel are registered in the binary running inthis process. Note that if you are loading a saved graph which used ops from tf.contrib, accessing (e.g.) tf.contrib.resampler should be done beforeimporting the graph, as contrib ops are lazily registered when the module is first accessed.

and I have read through all Issues about this problem, but just not solved yet.