This is an open source project (formerly named Listen, Attend and Spell - PyTorch Implementation) for end-to-end ASR implemented with Pytorch, the well known deep learning toolkit.
MIT License
1.18k
stars
318
forks
source link
If anyone trained this models,Please post librispeech WER and/or timit PER. #38
26.9 PER in test set, but I actually find a problem in dev set: PER is so high(greater than 1), and I'm trying to change my configration.