I want to set up a project using my own audio files and language model, but keeping the acoustic model from Tedlium project.
First two decodings have completed successfully (mono and tri1), but I get this error when I try to run tri2 and tri3:
gmm-latgen-faster-parallel --num-threads=4 --max-active=2000 --beam=10.0 --lattice-beam=6.0 --acoustic-scale=0.083333 --allow-partial=true --word-symbol-table=exp/tri3/graph/words.txt exp/tri3/final.mdl exp/tri3/graph/HCLG.fst 'ark,s,cs:apply-cmvn --utt2spk=ark:data/test/split1/1/utt2spk scp:data/test/split1/1/cmvn.scp scp:data/test/split1/1/feats.scp ark:- | add-deltas ark:- ark:- |' 'ark:|gzip -c > exp/tri3/decode_test.si/lat.1.gz'
apply-cmvn --utt2spk=ark:data/test/split1/1/utt2spk scp:data/test/split1/1/cmvn.scp scp:data/test/split1/1/feats.scp ark:-
add-deltas ark:- ark:-
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
terminate called after throwing an instance of 'std::runtime_error'
what(): ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
I want to set up a project using my own audio files and language model,
but keeping the acoustic model from Tedlium project.
First two decodings have completed successfully (mono and tri1), but I get
this error when I try to run tri2 and tri3:
gmm-latgen-faster-parallel --num-threads=4 --max-active=2000 --beam=10.0
--lattice-beam=6.0 --acoustic-scale=0.083333 --allow-partial=true
--word-symbol-table=exp/tri3/graph/words.txt exp/tri3/final.mdl
exp/tri3/graph/HCLG.fst 'ark,s,cs:apply-cmvn
--utt2spk=ark:data/test/split1/1/utt2spk scp:data/test/split1/1/cmvn.scp
scp:data/test/split1/1/feats.scp ark:- | add-deltas ark:- ark:- |'
'ark:|gzip -c > exp/tri3/decode_test.si/lat.1.gz'
apply-cmvn --utt2spk=ark:data/test/split1/1/utt2spk
scp:data/test/split1/1/cmvn.scp scp:data/test/split1/1/feats.scp ark:-
add-deltas ark:- ark:-
ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
terminate called after throwing an instance of 'std::runtime_error'
what(): ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
Actually, since the data dim is 39 and the model dim is 40, I think it's
most likely that he is using delta+accel features but the model expected
LDA+MLLT features.
Lucian, have a look at the decoding command line that was used when you
decoded your original data where you trained the model. Most likely it
will have a splicing step (splice-feats) followed by a projection step
(transform-feats)- you should try to replicate that when you decode.
If you are using a decoding script, it will automatically pick up the
LDA+MLLT features if you copy the final.mat to the directory where your
final.mdl exists. Also make sure to copy splice_opts and cmvn_opts to that
directory.
On Wed, Feb 18, 2015 at 8:03 AM, Lucian Georgescu lucgeo92@users.sf.net
wrote:
Hi,
I want to set up a project using my own audio files and language model,
but keeping the acoustic model from Tedlium project.
First two decodings have completed successfully (mono and tri1), but I get
this error when I try to run tri2 and tri3:
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
terminate called after throwing an instance of 'std::runtime_error'
what(): ERROR
(gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50)
Dim mismatch: data dim = 39 vs. model dim = 40
Hi,
I want to set up a project using my own audio files and language model, but keeping the acoustic model from Tedlium project.
First two decodings have completed successfully (mono and tri1), but I get this error when I try to run tri2 and tri3:
gmm-latgen-faster-parallel --num-threads=4 --max-active=2000 --beam=10.0 --lattice-beam=6.0 --acoustic-scale=0.083333 --allow-partial=true --word-symbol-table=exp/tri3/graph/words.txt exp/tri3/final.mdl exp/tri3/graph/HCLG.fst 'ark,s,cs:apply-cmvn --utt2spk=ark:data/test/split1/1/utt2spk scp:data/test/split1/1/cmvn.scp scp:data/test/split1/1/feats.scp ark:- | add-deltas ark:- ark:- |' 'ark:|gzip -c > exp/tri3/decode_test.si/lat.1.gz'
apply-cmvn --utt2spk=ark:data/test/split1/1/utt2spk scp:data/test/split1/1/cmvn.scp scp:data/test/split1/1/feats.scp ark:-
add-deltas ark:- ark:-
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
terminate called after throwing an instance of 'std::runtime_error'
what(): ERROR (gmm-latgen-faster-parallel:LogLikelihoodZeroBased():decodable-am-diag-gmm.cc:50) Dim mismatch: data dim = 39 vs. model dim = 40
[stack trace: ]
kaldi::KaldiGetStackTrace()
kaldi::KaldiErrorMessage::~KaldiErrorMessage()
kaldi::DecodableAmDiagGmmUnmapped::LogLikelihoodZeroBased(int, int)
kaldi::DecodableAmDiagGmmScaled::LogLikelihood(int, int)
kaldi::LatticeFasterDecoder::ProcessEmitting(kaldi::DecodableInterface)
kaldi::LatticeFasterDecoder::Decode(kaldi::DecodableInterface)
kaldi::DecodeUtteranceLatticeFasterClass::operator()()
kaldi::TaskSequencer<kaldi::decodeutterancelatticefasterclass>::RunTask(void*)
/lib/x86_64-linux-gnu/libpthread.so.0(+0x7e9a) [0x7fefc1081e9a]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7fefc05902ed]</kaldi::decodeutterancelatticefasterclass>
I understand that my data dimension is different than model dimension. How can I manage this?
Thanks.
perhaps you forget to perform splicing?
y.
On Wed, Feb 18, 2015 at 8:03 AM, Lucian Georgescu lucgeo92@users.sf.net
wrote:
Actually, since the data dim is 39 and the model dim is 40, I think it's
most likely that he is using delta+accel features but the model expected
LDA+MLLT features.
Lucian, have a look at the decoding command line that was used when you
decoded your original data where you trained the model. Most likely it
will have a splicing step (splice-feats) followed by a projection step
(transform-feats)- you should try to replicate that when you decode.
If you are using a decoding script, it will automatically pick up the
LDA+MLLT features if you copy the final.mat to the directory where your
final.mdl exists. Also make sure to copy splice_opts and cmvn_opts to that
directory.
Dan
On Wed, Feb 18, 2015 at 10:28 AM, Jan Trmal jtrmal@users.sf.net wrote:
I copied those files and now it works. Thank you!