Open itaipee opened 4 years ago
Your proposed fix seems fine. Could you do some basic testing and make a PR?
sure , is there any check-list for PR ? like regression or something i need to do beside the actual code change ? thanks , Itai
On Tue, Mar 17, 2020 at 4:30 AM Daniel Povey notifications@github.com wrote:
Your proposed fix seems fine. Could you do some basic testing and make a PR?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/kaldi-asr/kaldi/issues/3998#issuecomment-599845596, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALAMETEY2R4NVCVA7BNG2YLRH3OCXANCNFSM4LMMIT5Q .
-- Itai Peer
Not really.... You can run minilibrispeech (or yesno, if you want something extremely tiny) y.
On Wed, Mar 18, 2020 at 8:26 AM itaipee notifications@github.com wrote:
sure , is there any check-list for PR ? like regression or something i need to do beside the actual code change ? thanks , Itai
On Tue, Mar 17, 2020 at 4:30 AM Daniel Povey notifications@github.com wrote:
Your proposed fix seems fine. Could you do some basic testing and make a PR?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/kaldi-asr/kaldi/issues/3998#issuecomment-599845596, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ALAMETEY2R4NVCVA7BNG2YLRH3OCXANCNFSM4LMMIT5Q
.
-- Itai Peer
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kaldi-asr/kaldi/issues/3998#issuecomment-600592573, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACUKYXZY4G7V7MI44MLRLCTRIC4V5ANCNFSM4LMMIT5Q .
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
in case the last argument is variable name ,parse_opts might get stuck in infinite lop , for example "./steps/nnet2/train_pnorm.sh --cmd run.pl --lda-opts"
the problem is that "shift 2" command fails and the "--lda-opts" is still first argument $1.
Suggestion for fix
! shift 2 && echo "$0: no option for $name" 1>&2 && exit 1;
Since this is very basic script that is used everywhere , I'm not sure it is wise to change it
https://github.com/kaldi-asr/kaldi/blob/53c4bd19f8a9f5f2ec7280b00a0df8720b692cd9/egs/wsj/s5/utils/parse_options.sh#L85