module 'tensorflow' has no attribute 'to_float'

[shuiruge]

Description

TensorFlow 2.0 has no attribute to_float.

Environment information

OS: Windows 10 Home Edition

$ pip freeze | grep tensor
mesh-tensorflow==0.1.4
tensor2tensor==1.14.1
tensorboard==2.0.1
tensorflow==2.0.0
tensorflow-datasets==1.3.0
tensorflow-estimator==2.0.1
tensorflow-gan==2.0.0
tensorflow-hub==0.7.0
tensorflow-metadata==0.15.0
tensorflow-probability==0.7.0

$ python -V
Python 3.7.5

For bugs: reproduction and error logs

# Steps to reproduce:
$ conda create -n test python=3.7
$ conda activate test
$ pip install tensor2tensor[tensorflow]
$ python -c "from tensor2tensor.models.transformer import Transformer"

# Error logs:
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\me\Anaconda3\envs\nlp\lib\site-packages\tensor2tensor\models\__init__.py", line 25, in <module>
    from tensor2tensor.layers import modalities  # pylint: disable=g-import-not-at-top
  File "C:\Users\me\Anaconda3\envs\nlp\lib\site-packages\tensor2tensor\layers\modalities.py", line 28, in <module>
    from tensor2tensor.layers import common_attention
  File "C:\Users\me\Anaconda3\envs\nlp\lib\site-packages\tensor2tensor\layers\common_attention.py", line 954, in <module>
    def attention_bias_to_padding(attention_bias, cast_fn=tf.to_float):
AttributeError: module 'tensorflow' has no attribute 'to_float'

[arneyjfs]

I am new to tensorflow and just wanted to get started with some of the models. I also ran into this issue with the 'Quickstart' problem in the tensor2tensor documentation. Briefly switching back to tensorflow 1.15 has allowed me to get around this error for now so i can at least get to grips with the core functionality - keen to find a solution to allow tf2.0 if anyone has any ideas

[irismasha]

You need to change tf.to_float call to tf.cast(..., dtype=tf.float32)

[skr3178]

I tried the above solution: I get the following error

---

ValueError Traceback (most recent call last) in () 2 3 from tensor2tensor.utils import registry ----> 4 from tensor2tensor import problems 5 6 problems.available() #Show all problems

9 frames /usr/local/lib/python3.7/dist-packages/tensor2tensor/problems.py in () 31 32 ---> 33 all_problems.import_modules(all_problems.ALL_MODULES)

/usr/local/lib/python3.7/dist-packages/tensor2tensor/data_generators/all_problems.py in import_modules(modules) 138 for module in modules: 139 try: --> 140 importlib.import_module(module) 141 except ImportError as error: 142 errors.append((module, error))

/usr/lib/python3.7/importlib/init.py in import_module(name, package) 125 break 126 level += 1 --> 127 return _bootstrap._gcd_import(name[level:], package, level) 128 129

/usr/local/lib/python3.7/dist-packages/tensor2tensor/data_generators/algorithmic.py in () 24 from six.moves import range # pylint: disable=redefined-builtin 25 from tensor2tensor.data_generators import generator_utils as utils ---> 26 from tensor2tensor.data_generators import problem 27 from tensor2tensor.data_generators import text_encoder 28 from tensor2tensor.layers import modalities

/usr/local/lib/python3.7/dist-packages/tensor2tensor/data_generators/problem.py in () 29 from tensor2tensor.utils import data_reader 30 from tensor2tensor.utils import hparam ---> 31 from tensor2tensor.utils import metrics 32 from tensor2tensor.utils import mlperf_log 33

/usr/local/lib/python3.7/dist-packages/tensor2tensor/utils/metrics.py in () 23 24 from tensor2tensor.layers import common_layers ---> 25 from tensor2tensor.layers import modalities 26 from tensor2tensor.utils import bleu_hook 27 from tensor2tensor.utils import rouge

/usr/local/lib/python3.7/dist-packages/tensor2tensor/layers/modalities.py in () 26 from six.moves import range # pylint: disable=redefined-builtin 27 ---> 28 from tensor2tensor.layers import common_attention 29 from tensor2tensor.layers import common_audio 30 from tensor2tensor.layers import common_image_attention as cia

/usr/local/lib/python3.7/dist-packages/tensor2tensor/layers/common_attention.py in () 952 953 @expert_utils.add_name_scope() --> 954 def attention_bias_to_padding(attention_bias, cast_fn=tf.cast(..., dtype=tf.float32)): 955 """Inverse of attention_bias_ignore_padding(). 956

/usr/local/lib/python3.7/dist-packages/tensorflow/python/util/traceback_utils.py in error_handler(*args, **kwargs) 151 except Exception as e: 152 filtered_tb = _process_traceback_frames(e.traceback) --> 153 raise e.with_traceback(filtered_tb) from None 154 finally: 155 del filtered_tb

/usr/local/lib/python3.7/dist-packages/tensorflow/python/framework/constant_op.py in convert_to_eager_tensor(value, ctx, dtype) 100 dtype = dtypes.as_dtype(dtype).as_datatype_enum 101 ctx.ensure_initialized() --> 102 return ops.EagerTensor(value, ctx.device_name, dtype) 103 104

ValueError: Attempt to convert a value (Ellipsis) with an unsupported type (<class 'ellipsis'>) to a Tensor.