Tacotron-2: Synthesis using multiple gpus

[marzieh-razavi]

Hello,
I get the following error when trying to do the synthesis with multiple gpus:
ValueError: Variable Tacotron_model/inference/inputs_embedding already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarScope? Originally defined at:
tacotron/models/tacotron.py", line 112, in initialize
'inputs_embedding', [len(symbols), hp.embedding_dim], dtype=tf.float32)
Is the synthesis with multiple gpus not supported?
Thanks a lot for your help.

[n5-suzuki]

@marzieh-razavi Hello.I also got the same problem.
I do not know whether it is correct, but I modified it as follows.

line 25 of tacotron/synthesizer.py
before:

with tf.variable_scope('Tacotron_model') as scope:

after:

with tf.variable_scope('Tacotron_model', reuse=tf.AUTO_REUSE) as scope:

[marzieh-razavi]

@marzieh-razavi Hello.I also got the same problem.
I do not know whether it is correct, but I modified it as follows.

line 25 of tacotron/synthesizer.py
before:

with tf.variable_scope('Tacotron_model') as scope:

after:

with tf.variable_scope('Tacotron_model', reuse=tf.AUTO_REUSE) as scope:

Thank you very much