train_mmi.sh
6.56 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
#!/bin/bash
# Copyright 2012 Johns Hopkins University (Author: Daniel Povey). Apache 2.0.
# MMI training (or optionally boosted MMI, if you give the --boost option).
# 4 iterations (by default) of Extended Baum-Welch update.
#
# For the numerator we have a fixed alignment rather than a lattice--
# this actually follows from the way lattices are defined in Kaldi, which
# is to have a single path for each word (output-symbol) sequence.
# Begin configuration section.
cmd=run.pl
num_iters=4
boost=0.0
cancel=true # if true, cancel num and den counts on each frame.
zero_if_disjoint=false # if true, ignore stats from frames where num + den
# have no overlap.
tau=400
weight_tau=10
acwt=0.1
stage=0
# End configuration section
echo "$0 $@" # Print the command line for logging
[ -f ./path.sh ] && . ./path.sh; # source the path.
. parse_options.sh || exit 1;
if [ $# -ne 5 ]; then
echo "Usage: steps/train_mmi.sh <data> <lang> <ali> <denlats> <exp>"
echo " e.g.: steps/train_mmi.sh data/train_si84 data/lang exp/tri2b_ali_si84 exp/tri2b_denlats_si84 exp/tri2b_mmi"
echo "Main options (for others, see top of script file)"
echo " --boost <boost-weight> # (e.g. 0.1), for boosted MMI. (default 0)"
echo " --cancel (true|false) # cancel stats (true by default)"
echo " --cmd (utils/run.pl|utils/queue.pl <queue opts>) # how to run jobs."
echo " --config <config-file> # config containing options"
echo " --stage <stage> # stage to do partial re-run from."
echo " --tau # tau for i-smooth to last iter (default 200)"
exit 1;
fi
data=$1
lang=$2
alidir=$3
denlatdir=$4
dir=$5
mkdir -p $dir/log
for f in $data/feats.scp $alidir/{tree,final.mdl,ali.1.gz} $denlatdir/lat.1.gz; do
[ ! -f $f ] && echo "$0: no such file $f" && exit 1;
done
nj=`cat $alidir/num_jobs` || exit 1;
[ "$nj" -ne "`cat $denlatdir/num_jobs`" ] && \
echo "$alidir and $denlatdir have different num-jobs" && exit 1;
sdata=$data/split$nj
splice_opts=`cat $alidir/splice_opts 2>/dev/null`
mkdir -p $dir/log
cp $alidir/splice_opts $dir 2>/dev/null
[[ -d $sdata && $data/feats.scp -ot $sdata ]] || split_data.sh $data $nj || exit 1;
echo $nj > $dir/num_jobs
cp $alidir/tree $dir
cp $alidir/final.mdl $dir/0.mdl
silphonelist=`cat $lang/phones/silence.csl` || exit 1;
# Set up features
if [ -f $alidir/final.mat ]; then feat_type=lda; else feat_type=delta; fi
echo "$0: feature type is $feat_type"
case $feat_type in
delta) feats="ark,s,cs:apply-cmvn --norm-vars=false --utt2spk=ark:$sdata/JOB/utt2spk scp:$sdata/JOB/cmvn.scp scp:$sdata/JOB/feats.scp ark:- | add-deltas ark:- ark:- |";;
lda) feats="ark,s,cs:apply-cmvn --norm-vars=false --utt2spk=ark:$sdata/JOB/utt2spk scp:$sdata/JOB/cmvn.scp scp:$sdata/JOB/feats.scp ark:- | splice-feats $splice_opts ark:- ark:- | transform-feats $alidir/final.mat ark:- ark:- |"
cp $alidir/final.mat $dir
;;
*) echo "Invalid feature type $feat_type" && exit 1;
esac
[ -f $alidir/trans.1 ] && echo Using transforms from $alidir && \
feats="$feats transform-feats --utt2spk=ark:$sdata/JOB/utt2spk ark,s,cs:$alidir/trans.JOB ark:- ark:- |"
lats="ark:gunzip -c $denlatdir/lat.JOB.gz|"
if [[ "$boost" != "0.0" && "$boost" != 0 ]]; then
lats="$lats lattice-boost-ali --b=$boost --silence-phones=$silphonelist $alidir/final.mdl ark:- 'ark,s,cs:gunzip -c $alidir/ali.JOB.gz|' ark:- |"
fi
x=0
while [ $x -lt $num_iters ]; do
echo "Iteration $x of MMI training"
# Note: the num and den states are accumulated at the same time, so we
# can cancel them per frame.
if [ $stage -le $x ]; then
$cmd JOB=1:$nj $dir/log/acc.$x.JOB.log \
gmm-rescore-lattice $dir/$x.mdl "$lats" "$feats" ark:- \| \
lattice-to-post --acoustic-scale=$acwt ark:- ark:- \| \
sum-post --zero-if-disjoint=$zero_if_disjoint --merge=$cancel --scale1=-1 \
ark:- "ark,s,cs:gunzip -c $alidir/ali.JOB.gz | ali-to-post ark:- ark:- |" ark:- \| \
gmm-acc-stats2 $dir/$x.mdl "$feats" ark,s,cs:- \
$dir/num_acc.$x.JOB.acc $dir/den_acc.$x.JOB.acc || exit 1;
n=`echo $dir/{num,den}_acc.$x.*.acc | wc -w`;
[ "$n" -ne $[$nj*2] ] && \
echo "Wrong number of MMI accumulators $n versus 2*$nj" && exit 1;
$cmd $dir/log/den_acc_sum.$x.log \
gmm-sum-accs $dir/den_acc.$x.acc $dir/den_acc.$x.*.acc || exit 1;
rm $dir/den_acc.$x.*.acc
$cmd $dir/log/num_acc_sum.$x.log \
gmm-sum-accs $dir/num_acc.$x.acc $dir/num_acc.$x.*.acc || exit 1;
rm $dir/num_acc.$x.*.acc
# note: this tau value is for smoothing towards model parameters, not
# as in the Boosted MMI paper, not towards the ML stats as in the earlier
# work on discriminative training (e.g. my thesis).
# You could use gmm-ismooth-stats to smooth to the ML stats, if you had
# them available [here they're not available if cancel=true].
$cmd $dir/log/update.$x.log \
gmm-est-gaussians-ebw --tau=$tau $dir/$x.mdl $dir/num_acc.$x.acc $dir/den_acc.$x.acc - \| \
gmm-est-weights-ebw --weight-tau=$weight_tau - $dir/num_acc.$x.acc $dir/den_acc.$x.acc $dir/$[$x+1].mdl || exit 1;
rm $dir/{den,num}_acc.$x.acc
fi
# Some diagnostics: the objective function progress and auxiliary-function
# improvement.
tail -n 50 $dir/log/acc.$x.*.log | perl -e '$acwt=shift @ARGV; while(<STDIN>) { if(m/gmm-acc-stats2.+Overall weighted acoustic likelihood per frame was (\S+) over (\S+) frames/) { $tot_aclike += $1*$2; $tot_frames1 += $2; } if(m|lattice-to-post.+Overall average log-like/frame is (\S+) over (\S+) frames. Average acoustic like/frame is (\S+)|) { $tot_den_lat_like += $1*$2; $tot_frames2 += $2; $tot_den_aclike += $3*$2; } } if (abs($tot_frames1 - $tot_frames2) > 0.01*($tot_frames1 + $tot_frames2)) { print STDERR "Frame-counts disagree $tot_frames1 versus $tot_frames2\n"; } $tot_den_lat_like /= $tot_frames2; $tot_den_aclike /= $tot_frames2; $tot_aclike *= ($acwt / $tot_frames1); $num_like = $tot_aclike + $tot_den_aclike; $per_frame_objf = $num_like - $tot_den_lat_like; print "$per_frame_objf $tot_frames1\n"; ' $acwt > $dir/tmpf
objf=`cat $dir/tmpf | awk '{print $1}'`;
nf=`cat $dir/tmpf | awk '{print $2}'`;
rm $dir/tmpf
impr=`grep -w Overall $dir/log/update.$x.log | awk '{x += $10*$12;} END{print x;}'`
impr=`perl -e "print ($impr*$acwt/$nf);"` # We multiply by acwt, and divide by $nf which is the "real" number of frames.
echo "Iteration $x: objf was $objf, MMI auxf change was $impr" | tee $dir/objf.$x.log
x=$[$x+1]
done
echo "MMI training finished"
rm $dir/final.mdl 2>/dev/null
ln -s $x.mdl $dir/final.mdl
exit 0;