ibab / tensorflow-wavenet

A TensorFlow implementation of DeepMind's WaveNet paper
MIT License
5.41k stars 1.3k forks source link

CTC instead of HMM #160

Open nakosung opened 7 years ago

nakosung commented 7 years ago

http://www.cs.toronto.edu/~graves/icml_2006.pdf https://www.tensorflow.org/versions/r0.11/api_docs/python/nn.html#conectionist-temporal-classification-ctc

We already have CTC implementation in tensorflow. So could we try local conditioning based on CTC?

lemonzi commented 7 years ago

I'd prefer to not use a separate HMM for the alignments, so yes, any alternative is welcome :)

On Fri, Oct 21, 2016, 21:17 Nako Sung notifications@github.com wrote:

http://www.cs.toronto.edu/~graves/icml_2006.pdf https://www.tensorflow.org/versions/r0.11/api_docs/python/nn.html#conectionist-temporal-classification-ctc

We already have CTC implementation in tensorflow. So could we try local conditioning based on CTC?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ibab/tensorflow-wavenet/issues/160, or mute the thread https://github.com/notifications/unsubscribe-auth/ADCF5iNrRWrPPlgoLvivm3Lp_QwvWDSjks5q2WQugaJpZM4Kdt6f .