yh1008 / speech-to-text

mixlingual speech recognition system; hybrid (GMM+NNet) model; Kaldi + Keras
http://llcao.net/cu-deeplearning17/project.html
70 stars 19 forks source link

WARNING: triphone decode #13

Open yh1008 opened 7 years ago

yh1008 commented 7 years ago

seen warning: optional-silence SIL is seen only 79.5256679676% of the time at utterance begin. This may not be optimal.

Command:

steps/decode.sh --nj 8 --cmd run.pl exp/tri1/graph data/test exp/tri1/decode

Outputs:

decode.sh: feature type is delta
steps/diagnostic/analyze_lats.sh --cmd run.pl exp/tri1/graph exp/tri1/decode
analyze_phone_length_stats.py: WARNING: optional-silence SIL is seen only 79.5256679676% of the time at utterance begin.  This may not be optimal.
steps/diagnostic/analyze_lats.sh: see stats in exp/tri1/decode/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(4,36,220) and mean=87.0
steps/diagnostic/analyze_lats.sh: see stats in exp/tri1/decode/log/analyze_lattice_depth_stats.log
Not scoring because local/score.sh does not exist or not executable.