cannot import name 'conditional' from 'hyperas.distributions'

[last2win]

i use other people's code:

from __future__ import print_function

from hyperopt import Trials, STATUS_OK, tpe
from keras.datasets import mnist
from keras.layers.core import Dense, Dropout, Activation
from keras.models import Sequential
from keras.utils import np_utils

from hyperas import optim
from hyperas.distributions import choice, uniform, conditional

and error occured:

Traceback (most recent call last):

  File "<ipython-input-1-77a51f029857>", line 10, in <module>
    from hyperas.distributions import choice, uniform, conditional

ImportError: cannot import name 'conditional' from 'hyperas.distributions' (C:\Users\peter\Anaconda3\lib\site-packages\hyperas\distributions.py)

i tryed:

pip uninstall hyperas
pip install git+https://github.com/maxpumperla/hyperas.git
pip uninstall hyperopt
pip install git+https://github.com/hyperopt/hyperopt.git

but the problem is still there.

[maxpumperla]

conditional is definitely gone from master and the last release (if I remember correctly). You can get rid of the error by removing all the conditional calls from "other people's code". :D

[last2win]

ok, i solved the problem.

[ljakupi]

@maxpumperla what's the alternative for conditional as it is removed?

E.g. what could be the solution to such cases: if conditional({{choice(['three', 'four'])}}) == 'four': ....

... well, one could be: if {{choice(['three', 'four'])}} == 'four': ...

[BetterTogetherX]

@zhang0peter ，How to solve it

[last2win]

@zhang0peter ，How to solve it

i delete conditional

[cklat]

is there any solution/alternative for conditional? I have basically the same code snippet that @ljakupi has posted for optimizing the number of layers of the trained network. However, since conditional was dropped, I'm wondering how to realize this pipeline (i.e. optimizing for number of layers)