Search

ICSLP 2000 Beijing, China, October 16-20 2000. ... ICSLP 2002 Denver, USA, September 17-20 2002.

12 May 2009: I try to train the hmm model with the new feature (the TARGETRATE=3D160000.0,frame shite is 16ms and WINDOWSIZE is 32ms).=20 =20 Besause of the HDecode in ... 20 But with these lattices generated by edited HDecode, the HMMIRest tool still reports errors

13 Mar 2009: I found a bug in HTKBook: --- config.tex 2006-12-07 20:09:07.000000000 0900 config.tex.fixed 2008-05-22 20:47:06.000000000 0900 @@ -377,7 377,7 @@ &

13 Mar 2009: /HMath.c 2007-07-18 11:20:54.000000000 0900. /HMath.c 2007-07-18 11:26:35.000000000 0900 @@ -1287,6 1287,7 @@ for (i=j1; i=n;i++)

13 Mar 2009: v 1.8 2007/09/18 12:20:45 zen Exp $ HMem 3.4 CUED 25/04/06 : $Id: HMem.c,v 1.5 2007/10/03 07:20:14 ... zen Exp $ HLabel 3.4 CUED 25/04/06 : $Id: HLabel.c,v 1.5 2007/09/18 12:20:43 zen Exp $ HMath 3.3 CUED

13 Mar 2009: HCompV -m -S train_mfcc_paths.scp -M hmm1.a proto.txt This how i have been using the HCompV tool.=20 A sample of the.

12 May 2009: In turn, there are more frames output than expected. From an hour long WAV file (01:08:20.44)=A0sampled at 22050 Hz, this produces 410976 "10ms" output frames, making the

29 Jan 2010: htk_3.4.1/HTKLib/HAdapt.c 2009-03-12 18:41:39.000000000 0000 htk_3.4.2/HTKLib/HAdapt.c 2009-11-22 15:20:53.083000000 0000 @@

1 Oct 2009: NOSIL/TYPE_GMM_= MODEL_ALL-COVKIND_DIAGC-MIXES_16/models/synthesized_neutral/proto/proto=20 HTK Configuration Parameters[13] Module/Tool Parameter Value # HIFREQ 8000 # LOFREQ 0 # ENORMALISE FALSE # NUMCEPS 26 # ... nosil [20-109] stack smashing detected

20 Apr 2009. Nickolay V. Shmyrev. As far as I remember the issue here is that times after multiplication aren't precisely aligned, so little floating point mismatch makes HMMIRest think