error while using indian eng mfa for (ipa)

  File "FastSpeech2/model/modules.py", line 129, in forward
    x = x + pitch_embedding
RuntimeError: The size of tensor a (32) must match the size of tensor b (96) at non-singleton dimension 1.

I got this error , while training a Indian English dataset 
Has someone tried training on some language which uses ipa ?
or has someone trained their own mfa?
#247 @ming024 @cyhuang-tw 