That's cool, and certainly cheaper than Google or Amazon's TTS feature
New tutorial on http://keras.io : build an automatic speech recognition system using a CTC loss and a model inspired by DeepSpeech2.
End-to-end (from raw wave files to inference with the trained model) in... *200 lines of code** 
https://keras.io/examples/audio/ctc_asr/ …pic.twitter.com/MdKNcU3ucF
-
-
-
It’s a STT pipeline not TTS
- Show replies
New conversation -
-
-
Cool! Did you try it with other languages? (ex italian)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I remember implementing this in keras in 2018! I had to use a custom layer for the ctcloss. It shows how far things have come, that it can be done in 200 lines
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
FINALLY
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
pls update docs for Audio Data in `code examples` section. (currently invisible)pic.twitter.com/qE9L4nEa6S
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.