@LGTrader I don't see any issue. Here is the gist.

Please make sure you have installed TF2.0. When I tried the above gist with TF1.13.1, it has thrown an error. So those four links work well for TF2.0 only. Please let me know what you think. If you agree, please close the issue. Thanks!

@jvishnuvardhan

I reworked the code and ran it in Jupyter Notebook on my system. I cannot download the same TF2.0 alpha version that you use because my CPU doesn't have an instruction set (AVX I think) that is now required so my version is built locally from the same code as of a couple of weeks ago. It does identify as 2.0-alpha-0

Also, a simple call to help(tf) actually does not show tf.optimizers as an option but does show tf.compat and tf.keras so I think everything is consistent.

For now I'm going to assume that possibly I caught the code in a weird state when I built this and will close the report. If I get a chance to rebuild newer code and still see it I think I can reopen it and we'll try again. If you think it's important to keep looking at this I assume you can reopen also?

Thanks!

import tensorflow as tf
print(tf.version)
opt1 = tf.compat.v2.keras.optimizers.Adagrad
opt2 = tf.compat.v2.optimizers.Adagrad
opt3 = tf.keras.optimizers.Adagrad
print(opt1)
print(opt2)
print(opt3)
2.0.0-alpha0
<class 'tensorflow.python.keras.optimizer_v2.adagrad.Adagrad'>
<class 'tensorflow.python.keras.optimizer_v2.adagrad.Adagrad'>
<class 'tensorflow.python.keras.optimizer_v2.adagrad.Adagrad'>
opt4 = tf.optimizers.Adagrad
print(opt4)

AttributeError Traceback (most recent call last)
in
----> 1 opt4 = tf.optimizers.Adagrad
2 print(opt4)

AttributeError: module 'tensorflow' has no attribute 'optimizers'

Help on package tensorflow:

NAME
tensorflow - Bring in all of the public TensorFlow interface into this module.

PACKAGE CONTENTS
_api (package)
app (package)
audio (package)
autograph (package)
bitwise (package)
compat (package)
compiler (package)
config (package)
contrib (package)
core (package)
data (package)
debugging (package)
distribute (package)
distributions (package)
dtypes (package)
errors (package)
estimator (package)
examples (package)
experimental (package)
feature_column (package)
gfile (package)
graph_util (package)
image (package)
initializers (package)
io (package)
keras (package)
layers (package)
libtensorflow_framework
linalg (package)
lite (package)
logging (package)
lookup (package)
losses (package)
manip (package)
math (package)
metrics (package)
nest (package)
nn (package)
profiler (package)
python (package)
python_io (package)
quantization (package)
queue (package)
ragged (package)
random (package)
raw_ops (package)
resource_loader (package)
saved_model (package)
sets (package)
signal (package)
sparse (package)
spectral (package)
strings (package)
summary (package)
sysconfig (package)
test (package)
tools (package)
tpu (package)
train (package)
user_ops (package)
v1
version (package)

@LGTrader I am reopening it as I noticed Attribute error with tf-nightly. I don't see any issue with TF2.0.0-alpha0. We need to check whether we need to update tf-nightly or the docs on website. Thanks for finding this issue.

AttributeError Traceback (most recent call last)
in ()
3 opt2 = tf.compat.v2.optimizers.Adagrad
4 opt3 = tf.keras.optimizers.Adagrad
----> 5 opt4 = tf.optimizers.Adagrad

/usr/local/lib/python3.6/dist-packages/tensorflow/python/util/deprecation_wrapper.py in getattr(self, name)
104 if name.startswith('dw'):
105 raise AttributeError('Accessing local variables before they are created.')
--> 106 attr = getattr(self._dw_wrapped_module, name)
107 if (self._dw_warning_count < _PER_MODULE_WARNING_LIMIT and
108 name not in self._dw_deprecated_printed):

AttributeError: module 'tensorflow' has no attribute 'optimizers'

Thanks. You saved me building again. (about 4 hours on my machine) I'll look for feedback in this thread and rebuild to verify fixes later if appropriate.

I think the latest change to enable tf.optimizers.Adagrad might not be in the alpha0 release, can you install nightly and try it? pip install tf-nightly-gpu-2.0-preview