op type not registered `NormalizeUTF8` initializing BERT

[jaismith]

Getting the following error when initializing a BERT model using this package:

FileNotFoundError: Op type not registered 'NormalizeUTF8' in binary running on <computer name>.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.
 If trying to load on a different device from the computational device, consider using setting the `experimental_io_device` option on tf.saved_model.LoadOptions to the io_device such as '/job:localhost'.

Here's the code initializing the model:

# use ALBERT
tfhub_handle_encoder = 'https://tfhub.dev/tensorflow/albert_en_base/2'
tfhub_handle_preprocess = 'https://tfhub.dev/tensorflow/albert_en_preprocess/3'

bert_preprocess_model = hub.KerasLayer(tfhub_handle_preprocess)
bert_model = hub.KerasLayer(tfhub_handle_encoder)

I've tried installing via the quick install script and the Conda instructions in #153, both throw the same error. Would appreciate if anyone had any insights.

---

Edit: modules are imported as follows:

import tensorflow as tf
import tensorflow_hub as hub
import tensorflow_addons.text as text
import tensorflow_addons.optimizers as optimizers

[bksaini078]

Hello @jaismith ,
I have tried the same but was getting the same error message.
Even transformers from Huggingface also not working.