Timeshrinking
From SpeechWiki
(Difference between revisions)
Line 3: | Line 3: | ||
{| class="wikitable" | {| class="wikitable" | ||
! train <math>\tau</math> !! test <math>\tau</math> !! dev 2k utt WER !! frames dropped | ! train <math>\tau</math> !! test <math>\tau</math> !! dev 2k utt WER !! frames dropped | ||
+ | |- | ||
+ | | 1 (old baseline)|| 1 || [{{SERVER}}/speech/akantor/fisher/exp/triphone/test2kUttOnConvGaus.noUnits/config73/LATEST.log 51.6%] || 0% | ||
|- | |- | ||
| .6 || .6 || 69.3% || ~35% | | .6 || .6 || 69.3% || ~35% | ||
Line 8: | Line 10: | ||
| .9 || .9 || [{{SERVER}}/speech/akantor/fisher/exp/timeshrink/test/unit.tri.timeshrink.point9/LATEST.log 56.3%] || ~5% | | .9 || .9 || [{{SERVER}}/speech/akantor/fisher/exp/timeshrink/test/unit.tri.timeshrink.point9/LATEST.log 56.3%] || ~5% | ||
|- | |- | ||
- | | 1 | + | | 1 || .9 || [{{SERVER}}/speech/akantor/fisher/exp/timeshrink/testOnBaseline/unit.tri.timeshrink.point9/LATEST.log 53.9%]|| ~5% |
|- | |- | ||
- | | .9 || 1 || | + | | .9 || 1 || TE || ~5% |
|} | |} | ||
- | I have to check for bugs. It could be that the threshold is too low or it could be something else too. We should probably rerun baseline too, just to make sure I didn't optimize it unfairly | + | I have to check for bugs. It could be that the threshold is too low or it could be something else too. We should probably rerun baseline too, just to make sure I didn't optimize it unfairly. |
Revision as of 23:39, 15 September 2009
Fisher experiments
train <math>\tau</math> | test <math>\tau</math> | dev 2k utt WER | frames dropped |
---|---|---|---|
1 (old baseline) | 1 | 51.6% | 0% |
.6 | .6 | 69.3% | ~35% |
.9 | .9 | 56.3% | ~5% |
1 | .9 | 53.9% | ~5% |
.9 | 1 | TE | ~5% |
I have to check for bugs. It could be that the threshold is too low or it could be something else too. We should probably rerun baseline too, just to make sure I didn't optimize it unfairly.
Future Directions
- Can be viewed as a two-mode special case of best-first viterbi search. So make a real best-first lattice search. Mark mentioned some attempts in the 80'ies to do this.