Yajie Miao

EESEN's current multi-gpu implementation is the script steps/train_ctc_parallel_h.sh, which is based on naive model averaging. It is not stable yet. Some people are working on this from different aspects, but...

Parallel training from the very beginning is risky. Training with one GPU for one or two iterations and then switching to parallel training might give you greater stability. @gowayyed may...

if they truly don't exist in your training data, you can safely delete them but caution that by default, Eesen maps OOV words in your training transcripts to

An even simpler way is to generate forced alignment with the GMM-HMM, and take the CD states as CI CTC labels. With this, there is no need to consider context...

first you need to make sure that your cuda library has been installed by default, it is normally at /usr/local/ on linux

there is no problem with it, this is normal did you see problems with the subsequent training?