Error: vectorfst:: read: read failed <unspecfied>

761 views
Skip to first unread message

Kina T.

unread,
Dec 2, 2017, 9:39:55 AM12/2/17
to kaldi-help
After the train_mono.sh script is successfully run, the following is error occurred. I try to read about the error from the kaldi help and i got related with the memory problem. I have 8GB memory and i checked when i run the script gain, it conusmed arround 5GB during this time and it stay for a long time and came up with this error.  is it the memory problem or may be other issue?
Thanks.

===== monophone training =====

steps/train_mono.sh --boost-silence 1.25 --nj 20 --cmd run.pl data/train data/lang exp/mono
steps/train_mono.sh: Initializing monophone system.
steps/train_mono.sh: Compiling training graphs
steps/train_mono.sh: Aligning data equally (pass 0)
steps/train_mono.sh: Pass 1
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 2
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 3
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 4
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 5
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 6
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 7
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 8
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 9
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 10
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 11
steps/train_mono.sh: Pass 12
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 13
steps/train_mono.sh: Pass 14
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 15
steps/train_mono.sh: Pass 16
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 17
steps/train_mono.sh: Pass 18
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 19
steps/train_mono.sh: Pass 20
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 21
steps/train_mono.sh: Pass 22
steps/train_mono.sh: Pass 23
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 24
steps/train_mono.sh: Pass 25
steps/train_mono.sh: Pass 26
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 27
steps/train_mono.sh: Pass 28
steps/train_mono.sh: Pass 29
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 30
steps/train_mono.sh: Pass 31
steps/train_mono.sh: Pass 32
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 33
steps/train_mono.sh: Pass 34
steps/train_mono.sh: Pass 35
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 36
steps/train_mono.sh: Pass 37
steps/train_mono.sh: Pass 38
steps/train_mono.sh: Aligning data
steps/train_mono.sh: Pass 39
steps/diagnostic/analyze_alignments.sh --cmd run.pl data/lang exp/mono
steps/diagnostic/analyze_alignments.sh: see stats in exp/mono/log/analyze_alignments.log
586 warnings in exp/mono/log/acc.*.*.log
1837 warnings in exp/mono/log/update.*.log
7050 warnings in exp/mono/log/align.*.*.log
exp/mono: nj=20 align prob=-97.87 over 25.99h [retry=0.1%, fail=0.0%] states=647 gauss=982
steps/train_mono.sh: Done training monophone system in exp/mono

===== monophone decoding =====

WARNING: the --mono, --left-biphone and --quinphone options are now deprecated and ignored.
tree-info exp/mono/tree
tree-info exp/mono/tree
fsttablecompose data/lang/L_disambig.fst data/lang/G.fst
fstdeterminizestar --use-log=true
fstminimizeencoded
fstpushspecial
ERROR: VectorFst::Read: Read failed: <unspecified>
ERROR (fstdeterminizestar[5.2.204~1-08848]:ReadFstKaldi():kaldi-fst-io.cc:40) Could not read fst from standard input

[ Stack-Trace: ]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)
kaldi::MessageLogger::~MessageLogger()
fst::ReadFstKaldi(std::string)
main
__libc_start_main
fstdeterminizestar() [0x4e1ce9]

ERROR: FstHeader::Read: Bad FST header: -
ERROR (fstminimizeencoded[5.2.204~1-08848]:ReadFstKaldi():kaldi-fst-io.cc:35) Reading FST: error reading FST header from standard input

[ Stack-Trace: ]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)
kaldi::MessageLogger::~MessageLogger()
fst::ReadFstKaldi(std::string)
main
__libc_start_main
fstminimizeencoded() [0x47e229]

ERROR: FstHeader::Read: Bad FST header: -
ERROR (fstpushspecial[5.2.204~1-08848]:ReadFstKaldi():kaldi-fst-io.cc:35) Reading FST: error reading FST header from standard input

[ Stack-Trace: ]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)
kaldi::MessageLogger::~MessageLogger()
fst::ReadFstKaldi(std::string)
main
__libc_start_main
fstpushspecial() [0x408869]

ERROR: FstHeader::Read: Bad FST header: standard input

Daniel Povey

unread,
Dec 2, 2017, 1:05:32 PM12/2/17
to kaldi-help
You will have to reduce the number of jobs it's decoding with (--nj option).


--
Go to http://kaldi-asr.org/forums.html find out how to join
---
You received this message because you are subscribed to the Google Groups "kaldi-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.
To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/6ac92519-55a6-44b7-8544-b2564c8792e1%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Jesus_Is_Lord

unread,
Dec 3, 2017, 12:35:00 PM12/3/17
to kaldi-help
I've got exactly the same type of error except for the result from train_mono.sh contains [retry=0.0%, fail=0.1%]:

user@mico-platform:~/1111/kaldi/kaldi-script$ ./04_train_mono.sh 

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh --nj 16 /home/user/1111/kaldi/data/train /home/user/1111/kaldi/lang /home/user/1111/kaldi/exp/System1/mono

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Initializing monophone system.

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Compiling training graphs

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data equally (pass 0)

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 1

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 2

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 3

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 4

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 5

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 6

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 7

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 8

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 9

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 10

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 11

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 12

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 13

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 14

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 15

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 16

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 17

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 18

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 19

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 20

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 21

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 22

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 23

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 24

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 25

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 26

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 27

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 28

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 29

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 30

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 31

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 32

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 33

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 34

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 35

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 36

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 37

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 38

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Aligning data

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Pass 39

steps/diagnostic/analyze_alignments.sh --cmd run.pl /home/user/1111/kaldi/lang /home/user/1111/kaldi/exp/System1/mono

steps/diagnostic/analyze_alignments.sh: see stats in /home/user/1111/kaldi/exp/System1/mono/log/analyze_alignments.log

6260 warnings in /home/user/1111/kaldi/exp/System1/mono/log/acc.*.*.log

4 warnings in /home/user/1111/kaldi/exp/System1/mono/log/compile_graphs.*.log

2760 warnings in /home/user/1111/kaldi/exp/System1/mono/log/update.*.log

6376 warnings in /home/user/1111/kaldi/exp/System1/mono/log/align.*.*.log

/home/user/1111/kaldi/exp/System1/mono: nj=16 align prob=-95.03 over 19.95h [retry=0.0%, fail=0.1%] states=737 gauss=977

/home/user/1111/kaldi-trunk/egs/wsj/s5/steps/train_mono.sh: Done training monophone system in /home/user/1111/kaldi/exp/System1/mono

WARNING: the --mono, --left-biphone and --quinphone options are now deprecated and ignored.

tree-info /home/user/1111/kaldi/exp/System1/mono/tree 

tree-info /home/user/1111/kaldi/exp/System1/mono/tree 

make-h-transducer --disambig-syms-out=/home/user/1111/kaldi/exp/System1/mono/graph/disambig_tid.int --transition-scale=1.0 /home/user/1111/kaldi/lang/tmp/ilabels_1_0 /home/user/1111/kaldi/exp/System1/mono/tree /home/user/1111/kaldi/exp/System1/mono/final.mdl 

fstrmsymbols /home/user/1111/kaldi/exp/System1/mono/graph/disambig_tid.int 

fstminimizeencoded 

fstrmepslocal 

fstdeterminizestar --use-log=true 

fsttablecompose /home/user/1111/kaldi/exp/System1/mono/graph/Ha.fst /home/user/1111/kaldi/lang/tmp/CLG_1_0.fst 

ERROR: VectorFst::Read: Read failed: <unspecified>

ERROR (fstdeterminizestar[5.2.196~1-8569f]:ReadFstKaldi():kaldi-fst-io.cc:40) Could not read fst from standard input


[ Stack-Trace: ]

fstdeterminizestar() [0x65717a]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)

kaldi::MessageLogger::~MessageLogger()

fst::ReadFstKaldi(std::string)

main

__libc_start_main

fstdeterminizestar() [0x55f589]


ERROR: FstHeader::Read: Bad FST header: -

ERROR (fstrmsymbols[5.2.196~1-8569f]:ReadFstKaldiGeneric():kaldi-fst-io.cc:53) Reading FST: error reading FST header from standard input


[ Stack-Trace: ]

fstrmsymbols() [0x57615e]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)

kaldi::MessageLogger::~MessageLogger()

fst::ReadFstKaldiGeneric(std::string, bool)

main

__libc_start_main

fstrmsymbols() [0x4b9cb9]


ERROR: FstHeader::Read: Bad FST header: -

ERROR (fstrmepslocal[5.2.196~1-8569f]:ReadFstKaldi():kaldi-fst-io.cc:35) Reading FST: error reading FST header from standard input


[ Stack-Trace: ]

fstrmepslocal() [0x59daee]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)

kaldi::MessageLogger::~MessageLogger()

fst::ReadFstKaldi(std::string)

main

__libc_start_main

fstrmepslocal() [0x4d5549]


ERROR: FstHeader::Read: Bad FST header: -

ERROR (fstminimizeencoded[5.2.196~1-8569f]:ReadFstKaldi():kaldi-fst-io.cc:35) Reading FST: error reading FST header from standard input


[ Stack-Trace: ]

fstminimizeencoded() [0x5f068c]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)

kaldi::MessageLogger::~MessageLogger()

fst::ReadFstKaldi(std::string)

main

__libc_start_main

fstminimizeencoded() [0x516399]


I've tried --nj with 10, 8 and 6, but the error is the same. Do you've any idea? Thanks in advance and kind regards,
To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+...@googlegroups.com.

Daniel Povey

unread,
Dec 3, 2017, 12:40:33 PM12/3/17
to kaldi-help
Hello Jesus_is_Lord,
I think it's a memory issue while building the decoding graph.  Maybe if you used a smaller language model it would help.  I don't know what scripts you are trying to run so I can't comment more specifically.
Dan



To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.

To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/2a071365-ab1d-4d3e-a438-4f96c2320ce6%40googlegroups.com.

Jesus_Is_Lord

unread,
Dec 3, 2017, 1:01:02 PM12/3/17
to kaldi-help
Hi Dan,
thanks for your quick response! I'm running kaldi scripts on a virtual box with 12GB RAM. Please get attached the script.
04_train_mono.sh

Daniel Povey

unread,
Dec 3, 2017, 1:07:03 PM12/3/17
to kaldi-help
It's failing somewhere in mkgraph.sh, there is a pipe where fsttablecompose is supposed to write to its standard output and that gets read by fstdeterminizestar.  Something is going wrong there, but it's not clear what.  I suspect an issue with your virtual machine (e.g. an OS-level issue) because I'm not seeing any printed messages about out-of-memory.  But breaking up that pipe using temporary files may help to debug it.
I can't help you with  how to do that though- you should find someone local who understands UNIX scripting if you can't do it yourself.

Dan


To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.

To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/b46b16d4-0851-40ef-8093-c8f1a2abb92d%40googlegroups.com.

Jesus_Is_Lord

unread,
Dec 8, 2017, 5:30:54 PM12/8/17
to kaldi-help
modifying the .../mkgraph.sh with:

  fsttablecompose $dir/Ha.fst "$clg" > tmpfsttablecompose

  cat tmpfsttablecompose | fstdeterminizestar --use-log=true > tmpfstdeterminizestar

  rm -f tmpfsttablecompose

  cat tmpfstdeterminizestar | fstrmsymbols $dir/disambig_tid.int > tmpfstrmsymbols

  rm tmpfstdeterminizestar

  cat tmpfstrmsymbols  | fstrmepslocal > tmpfstrmepslocal

  rm -f tmpfstrmsymbols

  cat tmpfstrmepslocal | fstminimizeencoded > $dir/HCLGa.fst.$$ || exit 1;

  rm -f tmpfstrmepslocal

  mv $dir/HCLGa.fst.$$ $dir/HCLGa.fst

  fstisstochastic $dir/HCLGa.fst || echo "HCLGa is not stochastic"


running on a machine with better RAM size has fixed our problem. Thanks a lot Dan for your suggestion. 
Message has been deleted

Daniel Povey

unread,
Dec 8, 2017, 6:28:25 PM12/8/17
to kaldi-help
It looks to me like you are using Yajie Miao's repository.
I am not going to support that.


On Fri, Dec 8, 2017 at 6:26 PM, Jesus_Is_Lord <yonasde...@gmail.com> wrote:
Hi Dan, I'm trying to generate a DNN based graph on the top of  04c_train_SAT_FMLLR.sh by adapting the script:
 https://github.com/besacier/ALFFA_PUBLIC/blob/f49a11472d849cb8162fa050a7555c31bcec2e60/ASR/WOLOF/kaldi-scripts/05_run_dnn.sh. I'm already done with generating fmllr features with 04c_train_SAT_FMLLR.sh, in the 05_run_dnn.sh script I need to adjust parameter settings for gmmdir=exp/system1_NoLengthContrast/tri3b (that can be replaced by exp/System1/tri3b), whereas I can't figure out how I set the parameter data_fmllr=data-fmllr/system1_NoLengthContrast, as I don't have anything with **data-fmllr**. Do you have any idea?

--
Go to http://kaldi-asr.org/forums.html find out how to join
---
You received this message because you are subscribed to the Google Groups "kaldi-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.
To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/0f034598-0e6a-4152-aec2-5a7f4a77e4bb%40googlegroups.com.

Daniel Povey

unread,
Dec 8, 2017, 6:30:44 PM12/8/17
to kaldi-help
Oh I see it's someone else's repo.
Anyway I won't support it myself.  It's probably an older nnet1-based recipe.

Jesus_Is_Lord

unread,
Dec 8, 2017, 6:32:24 PM12/8/17
to kaldi-help
Hi Dan, I'm trying to generate a DNN based graph on the top of  04c_train_SAT_FMLLR.sh by adapting the script:
 https://github.com/besacier/ALFFA_PUBLIC/blob/f49a11472d849cb8162fa050a7555c31bcec2e60/ASR/WOLOF/kaldi-scripts/05_run_dnn.sh. I'm already done with generating fmllr features with 04c_train_SAT_FMLLR.sh, in the 05_run_dnn.sh script I need to adjust parameter settings for gmmdir=exp/system1_NoLengthContrast/tri3b (I replace it by exp/System1/tri3b), whereas I can't figure out how to set replace the parameter data_fmllr=data-fmllr/system1_NoLengthContrast, as I don't have anything with **data-fmllr**. Do you have any idea?

Jesus_Is_Lord

unread,
Dec 8, 2017, 7:04:43 PM12/8/17
to kaldi-help
do you know any latest nnet based recipe that could help me make use of the existing fmllr features generated by 04c_train_SAT_FMLLR.sh and produce a DNN based graph?

Daniel Povey

unread,
Dec 8, 2017, 7:14:05 PM12/8/17
to kaldi-help
The latest recipes don't use fMLLR features, so no.


To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.

To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/361a226d-f7b8-42ee-9e8e-067bbd2a6ec6%40googlegroups.com.

Jesus_Is_Lord

unread,
Dec 8, 2017, 7:24:13 PM12/8/17
to kaldi-help
I've 04d_train_MMI_FMMI.sh, not run yet. Do you think it helps to use the latest recipes? If so please could suggest one?

Daniel Povey

unread,
Dec 8, 2017, 7:28:36 PM12/8/17
to kaldi-help
You are asking way too many questions.  Please don't email the list any  more today.
You can try the tedlium s5_r2 recipe, or the "yesno" recipe or the "mini_librispeech" recipe just to get used to the scripts.  But get them from the official Kaldi repository at https://github.com/kaldi-asr/kaldi.


To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.

To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/7ceb655a-bd36-4ffd-8a78-aefdb6a12c9b%40googlegroups.com.

Jesus_Is_Lord

unread,
Dec 10, 2017, 7:53:04 PM12/10/17
to kaldi-help
Here is the log error message:

WARNING (transform-feats[5.3.24~1-c948]:main():transform-feats.cc:87) No fMLLR transform available for utterance 12_d512023, producing no output for this utterance

LOG (transform-feats[5.3.24~1-c948]:main():transform-feats.cc:158) Overall average [pseudo-]logdet is -90.2113 over 14027 frames.

LOG (transform-feats[5.3.24~1-c948]:main():transform-feats.cc:161) Applied transform to 20 utterances; 0 had errors.

LOG (transform-feats[5.3.24~1-c948]:main():transform-feats.cc:161) Applied transform to 0 utterances; 20 had errors.

WARNING (copy-feats[5.3.24~1-c948]:Close():kaldi-io.cc:512) Pipe apply-cmvn  --utt2spk=ark:data/test/split18/10/utt2spk scp:data/test/split18/10/cmvn.scp scp:data/test/split18/10/feats.scp ark:- | splice$

ERROR (copy-feats[5.3.24~1-c948]:~SequentialTableReaderArchiveImpl():util/kaldi-table-inl.h:678) TableReader: error detected closing archive 'apply-cmvn  --utt2spk=ark:data/test/split18/10/utt2spk scp:da$


[ Stack-Trace: ]

copy-feats() [0x561e8c]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)

kaldi::MessageLogger::~MessageLogger()

kaldi::SequentialTableReaderArchiveImpl<kaldi::KaldiObjectHolder<kaldi::Matrix<float> > >::~SequentialTableReaderArchiveImpl()

kaldi::SequentialTableReaderArchiveImpl<kaldi::KaldiObjectHolder<kaldi::Matrix<float> > >::~SequentialTableReaderArchiveImpl()

kaldi::SequentialTableReader<kaldi::KaldiObjectHolder<kaldi::Matrix<float> > >::~SequentialTableReader()

main

__libc_start_main

copy-feats() [0x4a11a9]


Do you have any idea?

Daniel Povey

unread,
Dec 10, 2017, 7:56:42 PM12/10/17
to kaldi-help
It would most likely give that error because a previous stage had failed (or maybe because the disk was full or because you were simultaneously running more than one decoding job with output to the same directory).  I suspect it was the last one (running 2 identical jobs at the same time).



To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.

To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/a9ab39e4-72fe-451d-954e-03cc2400c905%40googlegroups.com.

Jesus_Is_Lord

unread,
Dec 10, 2017, 8:16:47 PM12/10/17
to kaldi-help
I've reduced the number of jobs to 1, but still have this error:

/home/user/1111/kaldi-trunk/kaldi/egs/wsj/s5/steps/nnet/make_fmllr_feats.sh --nj 1 --cmd /home/user/1111/kaldi-trunk/kaldi/egs/wsj/s5/utils/run.pl --transform-dir /home/user/1111/kaldi/exp/System1/tri3b /home/user/1111/kaldi/data-fmllr/System1/test data/test /home/user/1111/kaldi/exp/System1/tri3b /home/user/1111/kaldi/data-fmllr/System1/test/log /home/user/1111/kaldi/data-fmllr/System1/test/data

/home/user/1111/kaldi-trunk/kaldi/egs/wsj/s5/steps/nnet/make_fmllr_feats.sh: feature type is lda_fmllr

/home/user/1111/kaldi-trunk/kaldi/egs/wsj/s5/utils/copy_data_dir.sh: copied data from data/test to /home/user/1111/kaldi/data-fmllr/System1/test

Checking /home/user/1111/kaldi/data-fmllr/System1/test/text ...

--> reading /home/user/1111/kaldi/data-fmllr/System1/test/text

--> text seems to be UTF-8 or ASCII, checking whitespaces

--> text contains only allowed whitespaces

/home/user/1111/kaldi-trunk/kaldi/egs/wsj/s5/utils/validate_data_dir.sh: Successfully validated data-directory /home/user/1111/kaldi/data-fmllr/System1/test

bash: line 1:  2677 Aborted                 ( copy-feats '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:- | splice-feats --left-context=3 --right-context=3 ark:- ark:- | transform-feats /home/user/1111/kaldi/exp/System1/tri3b/final.mat ark:- ark:- | transform-feats --utt2spk=ark:data/test/split1/1/utt2spk "ark:cat /home/user/1111/kaldi/exp/System1/tri3b/trans.* |" ark:- ark:- |' ark,scp:/home/user/1111/kaldi/data-fmllr/System1/test/data/feats_fmllr_test.1.ark,/home/user/1111/kaldi/data-fmllr/System1/test/data/feats_fmllr_test.1.scp ) 2>> /home/user/1111/kaldi/data-fmllr/System1/test/log/make_fmllr_feats.1.log >> /home/user/1111/kaldi/data-fmllr/System1/test/log/make_fmllr_feats.1.log

run.pl: job failed, log is in /home/user/1111/kaldi/data-fmllr/System1/test/log/make_fmllr_feats.1.log

Daniel Povey

unread,
Dec 10, 2017, 8:19:11 PM12/10/17
to kaldi-help
I wasn't talking about the --nj parameter.  I was talking about if your had run decode.sh or whatever you are running, from two different locations.
I won't answer any more questions from you today.


To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.

To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/9fe8af6c-51d6-4050-a0c9-1b24fecb4a93%40googlegroups.com.
Message has been deleted

Jesus_Is_Lord

unread,
Dec 15, 2017, 11:53:50 AM12/15/17
to kaldi-help
Hi Dan,

Could you help me with this error from wsj/s5/steps/train_mmi_sgmm2.sh:
gzip: /home/user/1111/kaldi/exp/System1/sgmm2_5b2_mmi_b0.1_z/den_acc.0.1.gz: unexpected end of file
ERROR (sgmm2-sum-accs[5.3.24~1-c948]:Read():kaldi-matrix.cc:1613) Failed to read matrix from stream.   File position at start is -1, currently -1

[ Stack-Trace: ]
sgmm2-sum-accs() [0x5da59c]

kaldi::MessageLogger::HandleMessage(kaldi::LogMessageEnvelope const&, char const*)
kaldi::MessageLogger::~MessageLogger()
kaldi::Matrix<float>::Read(std::istream&, bool, bool)
kaldi::Matrix<double>::Read(std::istream&, bool, bool)
kaldi::Matrix<double>::Read(std::istream&, bool, bool)
kaldi::MleAmSgmm2Accs::Read(std::istream&, bool, bool)

Daniel Povey

unread,
Dec 15, 2017, 2:36:55 PM12/15/17
to kaldi-help
With regard to the copy command: there may be something strange with your version of 'cp'.  You should find someone local with linux experience to look into it because I think it's a bad use of my time.

The other error probably results from a previous error or possibly a full disk.  I don't have time to help further with this.


To unsubscribe from this group and stop receiving emails from it, send an email to kaldi-help+unsubscribe@googlegroups.com.

To post to this group, send email to kaldi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kaldi-help/f4182e28-a950-4fa5-9ab2-8d4dfa2f121b%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages