sequence: Changing priors update#32
Conversation
…AMI; a fix to aspire recipe (kaldi-asr#1314) * added reverberated TDNN+LSTM recipe on AMI; a fix to aspire recipe * added script version number
…xconfig descriptors (kaldi-asr#1338)
…tive training scripts
Conflicts: src/nnet3/nnet-general-component.cc
| nnet3-discriminative-copy-egs ark:$egs_dir/$egs_type.$egs_part.ark ark:- \| \ | ||
| nnet3-discriminative-subset-egs --srand=JOB --n=$prior_subset_size ark:- ark:- \| \ | ||
| nnet3-discriminative-merge-egs ark:- ark:- \| \ | ||
| nnet3-discriminative-merge-egs --minibatch-size=$minibatch_size ark:- ark:- \| \ |
There was a problem hiding this comment.
looks good generally...
what do you think about changing the $iter.adj.mdl to ${iter}_adj.mdl...
my concern is that with $iter.adj, it may not be clear to the user that the whole
string $iter.adj is treated as the iteration number for the purposes of the decoding script.
With an underscore separation, it's a little easier to see that they hang together.
… issue that Alexander Gorodetski pointed out on list RE a warning.
|
Thanks-- merging. |
|
Oh, wait, sorry... I've moved the main development to the 'shortcut' branch in the official Kaldi repo, this branch is dead now. |
vimalmanohar
force-pushed
the
shortcut
branch
from
cd6e0ae to
56a1970
Compare
14 participants
No description provided.