run_5a_clean_100.sh
2.23 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
#!/bin/bash
# This is p-norm neural net training, with the "fast" script, on top of adapted
# 40-dimensional features.
train_stage=-10
use_gpu=true
. ./cmd.sh
. ./path.sh
. utils/parse_options.sh
if $use_gpu; then
if ! cuda-compiled; then
cat <<EOF && exit 1
This script is intended to be used with GPUs but you have not compiled Kaldi with CUDA
If you want to use GPUs (and have them), go to src/, and configure and make on a machine
where "nvcc" is installed.
EOF
fi
parallel_opts="--gpu 1"
num_threads=1
minibatch_size=512
dir=exp/nnet5a_clean_100_gpu
else
# with just 4 jobs this might be a little slow.
num_threads=16
parallel_opts="--num-threads $num_threads"
minibatch_size=128
dir=exp/nnet5a_clean_100
fi
. ./cmd.sh
. utils/parse_options.sh
if [ ! -f $dir/final.mdl ]; then
if [[ $(hostname -f) == *.clsp.jhu.edu ]]; then
# spread the egs over various machines. will help reduce overload of any
# one machine.
utils/create_split_dir.pl /export/b0{1,2,3,4}/$USER/kaldi-data/egs/librispeech/s5/$dir/egs/storage $dir/egs/storage
fi
steps/nnet2/train_pnorm_fast.sh --stage $train_stage \
--samples-per-iter 400000 \
--parallel-opts "$parallel_opts" \
--num-threads "$num_threads" \
--minibatch-size "$minibatch_size" \
--num-jobs-nnet 4 --mix-up 8000 \
--initial-learning-rate 0.01 --final-learning-rate 0.001 \
--num-hidden-layers 4 \
--pnorm-input-dim 2000 --pnorm-output-dim 400 \
--cmd "$decode_cmd" \
data/train_clean_100 data/lang exp/tri4b_ali_clean_100 $dir || exit 1
fi
for test in test_clean test_other dev_clean dev_other; do
steps/nnet2/decode.sh --nj 20 --cmd "$decode_cmd" \
--transform-dir exp/tri4b/decode_tgsmall_$test \
exp/tri4b/graph_tgsmall data/$test $dir/decode_tgsmall_$test || exit 1;
steps/lmrescore.sh --cmd "$decode_cmd" data/lang_test_{tgsmall,tgmed} \
data/$test $dir/decode_{tgsmall,tgmed}_$test || exit 1;
steps/lmrescore_const_arpa.sh \
--cmd "$decode_cmd" data/lang_test_{tgsmall,tglarge} \
data/$test $dir/decode_{tgsmall,tglarge}_$test || exit 1;
steps/lmrescore_const_arpa.sh \
--cmd "$decode_cmd" data/lang_test_{tgsmall,fglarge} \
data/$test $dir/decode_{tgsmall,fglarge}_$test || exit 1;
done
exit 0;