search

domerin0 / rnn-speech

Character level speech recognizer using ctc loss with deep rnns in TensorFlow.
MIT License
77 stars 31 forks source link

Simpler management of starting point and global_step #37

Closed AMairesse closed 7 years ago

AMairesse commented 7 years ago

Training always start at the beginning of the training dataset, no more complicated use of global_step to try to continue from where it stopped.

New --start_from option for the user to be able to start at any point in the training dataset.