alumae / kaldi-offline-transcriber

Offline transcription system for Estonian using Kaldi
Other
226 stars 57 forks source link

final.mdl missing #1

Closed skpvox closed 10 years ago

skpvox commented 10 years ago

I'm receiving the following error when running it on the example audio file:

(Diarization output has been omitted).

Any ideas?

[...]
rm -rf build/audio/segmented/intervjuu201306211256
mkdir -p build/audio/segmented/intervjuu201306211256
cat build/diarization/intervjuu201306211256/show.seg | cut -f 3,4,8 -d " " | \
    while read LINE ; do \
        start=`echo $LINE | cut -f 1 -d " " | perl -npe '$_=$_/100.0'`; \
        len=`echo $LINE | cut -f 2 -d " " | perl -npe '$_=$_/100.0'`; \
        sp_id=`echo $LINE | cut -f 3 -d " "`; \
        timeformatted=`echo "$start $len" | perl -ne '@t=split(); $start=$t[0]; $len=$t[1]; $end=$start+$len; printf("%08.3f-%08.3f\n", $start,$end);'` ; \
        sox build/audio/base/intervjuu201306211256.wav --norm build/audio/segmented/intervjuu201306211256/intervjuu201306211256_${timeformatted}_${sp_id}.wav trim $start $len ; \
    done
sox WARN dither: dither clipped 1 samples; decrease volume?
sox WARN dither: dither clipped 1 samples; decrease volume?
sox WARN dither: dither clipped 1 samples; decrease volume?
sox WARN dither: dither clipped 1 samples; decrease volume?
mkdir -p `dirname build/trans/intervjuu201306211256/wav.scp`
/bin/ls build/audio/segmented/intervjuu201306211256/*.wav  | \
        perl -npe 'chomp; $orig=$_; s/.*\/(.*)_(\d+\.\d+-\d+\.\d+)_(S\d+)\.wav/\1-\3---\2/; $_=$_ .  " $orig\n";' | LC_ALL=C sort > build/trans/intervjuu201306211256/wav.scp
cat build/trans/intervjuu201306211256/wav.scp | perl -npe 's/\s+.*//; s/((.*)---.*)/\1 \2/' > build/trans/intervjuu201306211256/utt2spk
utils/utt2spk_to_spk2utt.pl build/trans/intervjuu201306211256/utt2spk > build/trans/intervjuu201306211256/spk2utt
rm -rf build/trans/intervjuu201306211256/mfcc
steps/make_mfcc.sh --mfcc-config conf/mfcc.conf --cmd "$train_cmd" --nj 1 \
        build/trans/intervjuu201306211256 build/trans/intervjuu201306211256/exp/make_mfcc build/trans/intervjuu201306211256/mfcc || exit 1
steps/make_mfcc.sh --mfcc-config conf/mfcc.conf --cmd run.pl --nj 1 build/trans/intervjuu201306211256 build/trans/intervjuu201306211256/exp/make_mfcc build/trans/intervjuu201306211256/mfcc
steps/make_mfcc.sh: [info]: no segments file exists: assuming wav.scp indexed by utterance.
Succeeded creating MFCC features for intervjuu201306211256
steps/compute_cmvn_stats.sh build/trans/intervjuu201306211256 build/trans/intervjuu201306211256/exp/make_mfcc build/trans/intervjuu201306211256/mfcc || exit 1;
steps/compute_cmvn_stats.sh build/trans/intervjuu201306211256 build/trans/intervjuu201306211256/exp/make_mfcc build/trans/intervjuu201306211256/mfcc
Succeeded creating CMVN stats for intervjuu201306211256
rm -rf build/trans/intervjuu201306211256/tri3b_mmi_pruned
mkdir -p build/trans/intervjuu201306211256/tri3b_mmi_pruned
(cd build/trans/intervjuu201306211256/tri3b_mmi_pruned; for f in ../../../fst/tri3b_mmi/*; do ln -s $f; done)
steps/decode_fmllr.sh --num-threads 10 --config conf/decode.conf --skip-scoring true --nj 1 --cmd "$decode_cmd" \
      --alignment-model build/fst/tri3b/final.alimdl --adapt-model build/fst/tri3b/final.mdl \
        build/fst/tri3b/graph_prunedlm build/trans/intervjuu201306211256 `dirname build/trans/intervjuu201306211256/tri3b_mmi_pruned/decode/log`
steps/decode_fmllr.sh --num-threads 10 --config conf/decode.conf --skip-scoring true --nj 1 --cmd run.pl --alignment-model build/fst/tri3b/final.alimdl --adapt-model build/fst/tri3b/final.mdl build/fst/tri3b/graph_prunedlm build/trans/intervjuu201306211256 build/trans/intervjuu201306211256/tri3b_mmi_pruned/decode
cat: build/trans/intervjuu201306211256/text: No such file or directory
steps/decode.sh --parallel-opts  --scoring-opts  --num-threads 10 --skip-scoring true --acwt 0.083333 --nj 1 --cmd run.pl --beam 10.0 --model build/fst/tri3b/final.alimdl --max-arcs -1 --max-active 2000 build/fst/tri3b/graph_prunedlm build/trans/intervjuu201306211256 build/trans/intervjuu201306211256/tri3b_mmi_pruned/decode.si
decode.sh: feature type is lda
steps/decode_fmllr.sh: no such file build/trans/intervjuu201306211256/tri3b_mmi_pruned/final.mdl
make: *** [build/trans/intervjuu201306211256/tri3b_mmi_pruned/decode/log] Error 1
alumae commented 10 years ago

Have you downloaded Estonian language and acoustic models, as explained in the README?