tensorflow.python.framework.errors_impl.NotFoundError: Op type not registered 'GatherTree' in binary running

[gyshi]

when i run gnmt with tf2.1, have this error . System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
• TensorFlow version and how it was installed (source or binary): tf2.1
• TensorFlow-Addons version and how it was installed (source or binary): 0.11
• Python version: 3.7
• Is GPU used? (yes/no): no

Describe the bug

Traceback (most recent call last):
  File "/home/shiguang/models/benchmarks/../models/language_translation/tensorflow/mlperf_gnmt/fp32/run_inference.py", line 90, in <module>
    importer.import_graph_def(graph_def, input_map={}, name="")
  File "/home/shiguang/.local/lib/python3.7/site-packages/tensorflow/python/util/deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "/home/shiguang/.local/lib/python3.7/site-packages/tensorflow/python/framework/importer.py", line 405, in import_graph_def
    producer_op_list=producer_op_list)
  File "/home/shiguang/.local/lib/python3.7/site-packages/tensorflow/python/framework/importer.py", line 497, in _import_graph_def_internal
    graph._c_graph, serialized, options)  # pylint: disable=protected-access
tensorflow.python.framework.errors_impl.NotFoundError: Op type not registered 'GatherTree' in binary running on    . 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.
Ran inference with batch size 32

A clear and concise description of what the bug is.

Code to reproduce the issue

Provide a reproducible test case that is the bare minimum necessary to generate the problem.

Other info / logs

Include any logs or source code that would be helpful to diagnose the problem. If including tracebacks, please include the full traceback. Large logs and files should be attached.

[bhack]

Do you have a minimal code that could be copy and pasted to reproduce this error?

[gyshi]

Do you have a minimal code that could be copy and pasted to reproduce this error? i use this toturial https://github.com/IntelAI/models/tree/master/benchmarks/language_translation/tensorflow/mlperf_gnmt because my tensorflow is also 2.1, but have internal upgrade, so i have to use more higher version for addons to build tensorflow-addons success. when i run this model, i will meet this error.

[seanpmorgan]

Hmmm TFA 0.11 hasn't yet been released are you sure you're correct about the versions. FWIW TFA 0.10 requires TensorFlow >= 2.2 which could be causing this error.

[gyshi]

TensorFlow

now, i will try other version, if i will meet same error, i will connect you, thx.

[gyshi]

Hmmm TFA 0.11 hasn't yet been released are you sure you're correct about the versions. FWIW TFA 0.10 requires TensorFlow >= 2.2 which could be causing this error.

i have try tfa 0.7, i meet the same error.

[guillaumekln]

The op GatherTree has been renamed to Addons>GatherTree in TensorFlow Addons 0.6. You should re-export the SavedModel with a newer version of TensorFlow Addons.

[gyshi]

The op GatherTree has been renamed to Addons>GatherTree in TensorFlow Addons 0.6. You should re-export the SavedModel with a newer version of TensorFlow Addons.

thx

[gyshi]

The op GatherTree has been renamed to Addons>GatherTree in TensorFlow Addons 0.6. You should re-export the SavedModel with a newer version of TensorFlow Addons.

if i want to rename Addons>GatherTree to GatherTree , I only need to modify REGISTER_OP("GatherTree")?

[gyshi]

The op GatherTree has been renamed to Addons>GatherTree in TensorFlow Addons 0.6. You should re-export the SavedModel with a newer version of TensorFlow Addons.

@guillaumekln if i want to rename Addons>GatherTree to GatherTree , I only need to modify REGISTER_OP("GatherTree")?

[guillaumekln]

This would require to rebuild the TensorFlow Addons package.This is not something we recommend.

Instead, you should either install TensorFlow Addons < 0.6 or open an issue on https://github.com/IntelAI/models to ask the maintainers to republish the trained model.

[gyshi]

This would require to rebuild the TensorFlow Addons package.This is not something we recommend.

Instead, you should either install TensorFlow Addons < 0.6 or open an issue on https://github.com/IntelAI/models to ask the maintainers to republish the trained model.

thx for your help.