Problem with sugartensor 1.0.0
mssmkmr opened this issue · comments
I am grateful your project the sugartensor.
Recently, I noticed speech-to-text-wavenet doesn't work with sugartensor 1.0.0.
train.py line 35
seq_len = tf.not_equal(x.sg_sum(dims=2), 0.).sg_int().sg_sum(dims=1)
is need to change to below.
seq_len = tf.not_equal(x.sg_sum(axis=2), 0.).sg_int().sg_sum(axis=1)
Thank you again for your project.
@mssmkmr I'm refactoring speech-to-text-wavenet. Just wait a couple of days please.
i am getting error while importing sugartensor : AttributeError: module 'tensorflow' has no attribute 'core'
is there another way than using sugartensor?