Description

I am using tensot2tensor which is built over Tensorflow and when I reach to the step pf training I got this error AttributeError: module 'tensorflow._api.v1.compat.v1.compat' has no attribute 'v1 and here is my code on google colab
...
!git clone https://github.com/tensorflow/tensor2tensor.git

import os
os.chdir('tensor2tensor')
!ls
!pip install .

import tensorflow as tf
import os
os.mkdir("/content/t2t")
os.mkdir("/content/t2t/data")
os.mkdir("/content/t2t/tmp")
os.mkdir("/content/t2t/train")
os.mkdir("/content/t2t/export")
os.mkdir("/content/t2t/translation")
os.mkdir("/content/t2t/event")
os.mkdir("/content/t2t/user")

DATA_DIR = "/content/t2t/data"
TMP_DIR = "/content/t2t/tmp"
TRAIN_DIR = "/content/t2t/train"
EXPORT_DIR = "/content/t2t/export"
TRANSLATIONS_DIR = "/content/t2t/translation"
EVENT_DIR = "/content/t2t/event"
USR_DIR = "/content/t2t/user"

PROBLEM = "translate_enfr_wmt_small8k" # We chose a problem translation English to French with 32.768 vocabulary
MODEL = "transformer" # Our model
HPARAMS = "transformer_big_single_gpu" # Hyperparameters for the model by default
# If you have a one gpu, use transformer_big_single_gpu

!t2t-datagen
--data_dir=$DATA_DIR
--tmp_dir=$TMP_DIR
--problem=$PROBLEM
--t2t_usr_dir=$USR_DIR

train_steps = 300000 # Total number of train steps for all Epochs
eval_steps = 100 # Number of steps to perform for each evaluation
batch_size = 4096
save_checkpoints_steps = 1000
ALPHA = 0.1
schedule = "continuous_train_and_eval"

!t2t-trainer
--data_dir=$DATA_DIR
--problem=$PROBLEM
--model=$MODEL
--hparams_set=$HPARAMS
--hparams="batch_size=$batch_size"
--schedule=$schedule
--output_dir=$TRAIN_DIR
--train_steps=$train_steps
--worker-gpu=1 \

OS: <your answer here>

$ pip freeze | grep tensor
# your output here

$ python -V
# your output here

For bugs: reproduction and error logs

# Steps to reproduce:
...

# Error logs:
...
AttributeError: in converted code:

    /usr/local/lib/python3.6/dist-packages/tensor2tensor/utils/t2t_model.py:323 call  *
        summarize_features(features, num_shards=self._num_datashards)
    /usr/local/lib/python3.6/dist-packages/tensor2tensor/utils/t2t_model.py:2240 summarize_features  *
        if not common_layers.should_generate_summaries():
    /usr/local/lib/python3.6/dist-packages/tensor2tensor/layers/common_layers.py:3494 should_generate_summaries  *
        if tf.compat.v1.get_variable_scope().reuse:
    /tensorflow-1.15.0/python3.6/tensorflow_core/python/util/module_wrapper.py:193 __getattr__
        attr = getattr(self._tfmw_wrapped_module, name)

    AttributeError: module 'tensorflow._api.v1.compat.v1.compat' has no attribute 'v1'