Replies: 4 comments
-
>>> georroussos |
Beta Was this translation helpful? Give feedback.
-
>>> Moester |
Beta Was this translation helpful? Give feedback.
-
>>> nmstoker |
Beta Was this translation helpful? Give feedback.
-
>>> Moester |
Beta Was this translation helpful? Give feedback.
-
>>> Moester
[July 3, 2020, 8:26am]
Here we are, training tacotron 2 has become this swift breeze...In
reality, leaving aside it is the route for the sweet spot in quality and
training curves, it hurts to say swift breeze a bit far from the truth.
At this point clearly no need to remind the scale of importance of
recent developments in TTS. Very fascinating to observe daily progress.
Still familiarizing what I am wondering:
[ slash #1] is LJ speech with transfer learning still the most
efficient for high grade results? slash
[ slash #2] how important is gradual training in expediting
training?
Is there an interesting point that's missing from this discussion and
considerations in the current state of things?
Cheers
[This is an archived TTS discussion thread from discourse.mozilla.org/t/are-we-now-headed-towards-the-brightest-future-in-training-efficiency]
Beta Was this translation helpful? Give feedback.
All reactions