marvin-nj

the same problem with you, so could anbody run clasy in ubuntu1804?

thanks! so , the expression feature can be decomposed if model trained sufficiently， right ?

> Hi @changtaoli > > Thanks for the report. IIRC, nothing about that function is changed, but let us look into it. > > I think you can suppress the...

Hello, Any improve with Wasserstein loss? I want to try it , please tell more ,thank you

I try tensorflow 1.5.0 with your modify sucessfully, thank you, guy! of course , correct cuda and cudnn is needed.

> Problem solved. Did not set the path to required files correctly. Hello, dear brother : I have the same question: I try english model sucessful, but chinese model failed....

> Which model are your referring to? A 16k chinese model trained by myself , accyrately model is ok, it decode sucessfully in kaldi

I fix it after reinstall my kaldiasr, everything is ok now , thank you !