(vt) [fuyongze@dnnd adapt]$ sh main.sh
DEBUG: output 1000147 ids out of 1000147 ids
/dnn4_added/fuyongze/yl/kaldi.voice_trigger/src/bin/copy-int-vector 'scp:head -n 1 temp/scp1 |' ark,t:-
LOG (copy-int-vector:main():copy-int-vector.cc:83) Copied 1 vectors of int32.
/dnn4_added/fuyongze/yl/kaldi.voice_trigger/src/bin/convert-ali-states temp/state.map scp:temp/scp1 ark,scp:temp/ark,temp/scp
LOG (convert-ali-states:main():convert-ali-states.cc:87) Succeeded converting alignments for 1000147 files, failed for 0
/dnn4_added/fuyongze/yl/kaldi.voice_trigger/src/bin/ali-to-post scp:temp/scp ark,scp:temp/ark.post,temp/scp.post
LOG (ali-to-post:main():ali-to-post.cc:65) Converted 1000147 alignments.
/dnn4_added/fuyongze/yl/kaldi.voice_trigger/src/bin/../featbin/copy-feats --htk-in scp:temp/trn.mfcc.lst.htk ark,scp:temp/feats.ark,temp/feats.scp
ERROR (copy-feats:ReadHtk():kaldi-matrix.cc:2036) Attempting to read HTK features from unsupported type (e.g. waveform or discrete features.
ERROR (copy-feats:ReadHtk():kaldi-matrix.cc:2036) Attempting to read HTK features from unsupported type (e.g. waveform or discrete features.
[stack trace: ]
kaldi::KaldiGetStackTrace()
kaldi::KaldiErrorMessage::~KaldiErrorMessage()
bool kaldi::ReadHtk<float>(std::istream&, kaldi::Matrix<float>*, kaldi::HtkHeader*)
kaldi::SequentialTableReaderScriptImpl<kaldi::HtkMatrixHolder>::LoadCurrent()
kaldi::SequentialTableReaderScriptImpl<kaldi::HtkMatrixHolder>::Value()
/dnn4_added/fuyongze/yl/kaldi.voice_trigger/src/bin/../featbin/copy-feats(main+0x93e) [0x40acfe]
/usr/lib64/libc.so.6(__libc_start_main+0xf5) [0x7fb5bde1e3d5]
/dnn4_added/fuyongze/yl/kaldi.voice_trigger/src/bin/../featbin/copy-feats() [0x40b84a]
/dnn4_added/fuyongze/yl/kaldi.voice_trigger/egs/wsj/s5/steps/nnet/train_scheduler.sh --train-tool nnet-train-frmshuff --use-gpu=0000:07:00.0 --objective-function=esx --feature-transform /dnn4_added/fuyongze/yl/vt/adapt/config/baseAM_313/cmvn-g_tr_splice10-2.nnet --learn-rate 0.001 --minibatch-size 512 /dnn4_added/fuyongze/yl/vt/adapt/config/baseAM_313/final.nnet ark:copy-feats scp:temp/feats.scp.rand.trn ark:- | ark:copy-feats scp:temp/feats.scp.rand.cv ark:- | scp:temp/scp.post scp:temp/scp.post temp/adapt
CROSSVAL PRERUN AVG.LOSS 1.0537 (ExtendedSoftmaxXent),
ITERATION 01: TRAIN AVG.LOSS 0.8173, (lrate0.001), CROSSVAL AVG.LOSS 0.9611, nnet accepted (final_iter01_learnrate0.001_tr0.8173_cv0.9611)
ITERATION 02:
这个错误是什么意思，影响网络结构生成吗？如何解决？