Skip to content

Conversation

@misbullah
Copy link
Owner

No description provided.

hhadian and others added 30 commits January 30, 2018 12:04
- make sure that lattice-generation, and subsequent search
  of 'oracle-transcript' in them uses the same data-split.
…rthonormalize-cleaned

Conflicts:
	egs/wsj/s5/steps/nnet3/chain/train.py
	src/nnet3/nnet-compile-looped.cc
	src/nnet3/nnet-simple-component.cc
…, OOV … (#2137)

* multi_en: Fixed acronym normalization, swbd lexicon preparation, OOV pronunciation generation, acoustic data sub-sampling,.etc; Added hub4_97 data
… 66 contributors in Kaldi. Thanks to Yishay Carmiel for gathering this info (#2214)
naxingyu and others added 26 commits June 25, 2018 01:11
.. avoids training failures due to NaNs appearing.
@misbullah misbullah closed this Jul 23, 2018
@misbullah misbullah changed the title Change print tot for 'rnnlm/get_num_splits.sh' Merge newest updated Jul 23, 2018
@misbullah misbullah reopened this Jul 23, 2018
@misbullah misbullah merged commit 9c9366b into misbullah:master Jul 23, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.