zeroqiaoba / ivector-xvector Goto Github PK
View Code? Open in Web Editor NEWExtract xvector and ivector under kaldi
Extract xvector and ivector under kaldi
When sign up in BaiduNetdisk I enter my phone as verification, yet no message is coming back to me. Please, can you upload the models in Google drive ?
Hello! May I ask what datasets you used for training this 3000h i-vector model?
Thanks a lot!
Thanks for sharing.
Is the model trained on Chinese or English audio ?
Can the model be used on a different language other than the one used for training ?
This question is probably related to #8
Thanks !
if the repo can support kaldi:gpu, the code will be more efficient.
During extracting ivectors, I encountered an error and here is the log information:
# compute-mfcc-feats --verbose=2 --config=conf/mfcc.conf scp,p:/data1/ivector-xvector/ivector/data/log/make_enrollmfcc/wav_data.1.scp ark:- | copy-feats --compress=true ark:- ark,scp:/data1/ivector-xvector/ivector/data/feat/mfcc/raw_mfcc_data.1.ark,/data1/ivector-xvector/ivector/data/feat/mfcc/raw_mfcc_data.1.scp
# Started at Fri Feb 15 15:04:00 CST 2019
#
bash: line 1: compute-mfcc-feats: command not found
bash: line 1: copy-feats: command not found
# Accounting: time=0 threads=1
# Ended (code 127) at Fri Feb 15 15:04:00 CST 2019, elapsed time 0 seconds
Is there someone who had seen the same problem? Any hint or suggestion will be appreciated.
when I execute ./enroll.sh wav/ 1 I get this error:
./enroll.sh: line 56: parse_options.sh: No such file or directory
steps/make_mfcc.sh --mfcc-config conf/mfcc.conf --nj 1 --cmd run.pl /content/ivector-xvector/xvector/data /content/ivector-xvector/xvector/data/log/make_enrollmfcc /content/ivector-xvector/xvector/data/feat/mfcc
ln: failed to create symbolic link './steps': File exists
ln: failed to create symbolic link './sid': File exists
ln: failed to create symbolic link './utils': File exists
utils/validate_data_dir.sh: WARNING: you have only one speaker. This probably a bad idea.
Search for the word 'bold' in http://kaldi-asr.org/doc/data_prep.html
for more information.
utils/validate_data_dir.sh: Successfully validated data-directory /content/ivector-xvector/xvector/data
steps/make_mfcc.sh: [info]: no segments file exists: assuming wav.scp indexed by utterance.
run.pl: job failed, log is in /content/ivector-xvector/xvector/data/log/make_enrollmfcc/make_mfcc_data.1.log****
Hi
Is there any other link were I can download the pre-trained i-vector models
It asks me to register and login in Baidunetdisk, and I dont have any account (I I dont understand chinese to do it)
Best
liang001.wen@node2:~/4.PoC/looking-to-listen/Xvector-ivector-master/xvector$ ./enroll.sh ./wav/ 1
1: without speaker
get utt2spk, DataPre start on Mon Jul 8 16:45:35 CST 2019
data preparatin finished successfully Mon Jul 8 16:45:35 CST 2019
sorted spk2utt ... : fix_data_dir start on Mon Jul 8 16:45:35 CST 2019
utils/fix_data_dir.sh: file /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data/utt2spk is not in sorted order or not unique, sorting it
utils/fix_data_dir.sh: file /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data/spk2utt is not in sorted order or not unique, sorting it
utils/fix_data_dir.sh: file /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data/wav.scp is not in sorted order or not unique, sorting it
fix_data_dir.sh: kept all 2 utterances.
fix_data_dir.sh: old files are kept in /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data/.backup
fix_data_dir finished successfully Mon Jul 8 16:45:35 CST 2019
FeatureForSpeaker start on Mon Jul 8 16:45:35 CST 2019
steps/make_mfcc.sh --mfcc-config conf/mfcc.conf --nj 1 --cmd utils/run.pl /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data/log/make_enrollmfcc /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data/feat/mfcc
utils/validate_data_dir.sh: Successfully validated data-directory /data/liang001.wen/4.PoC/looking-to-listen/Xvector-ivector-master/xvector/data
steps/make_mfcc.sh: [info]: no segments file exists: assuming wav.scp indexed by utterance.
usage: run.pl log-file command-line arguments... at utils/parallel/run.pl line 26.
When extracting x-vectors, the program get stucked at this. Could you please tell me how to do then?
The ivector pretrained model is not downloading in India. The BaiduNetdisk also doesn't open as it gives some error.
Could you please provide the downloaded version of the model in the repo. That would be of great help.
Thanks
I follow all the step and try to work with the flac example
but in /data/ folder the files are 0 bytes and no I didn't get the feat dir
fix_data_dir.sh: no utterances remained: not proceeding further.
bash enroll.sh ./wav/ 1
1: without speaker
get utt2spk, DataPre start on Thu Mar 14 02:58:15 PDT 2019
data preparatin finished successfully Thu Mar 14 02:58:16 PDT 2019
sorted spk2utt ... : fix_data_dir start on Thu Mar 14 02:58:16 PDT 2019
fix_data_dir.sh: kept all 5 utterances.
fix_data_dir.sh: old files are kept in /home/iseesaw/Desktop/Share/CrossDeviceSR/kaldi/ivector-xvector-master/xvector/data/.backup
fix_data_dir finished successfully Thu Mar 14 02:58:16 PDT 2019
FeatureForSpeaker start on Thu Mar 14 02:58:16 PDT 2019
steps/make_mfcc.sh --mfcc-config conf/mfcc.conf --nj 1 --cmd run.pl /home/iseesaw/Desktop/Share/CrossDeviceSR/kaldi/ivector-xvector-master/xvector/data /home/iseesaw/Desktop/Share/CrossDeviceSR/kaldi/ivector-xvector-master/xvector/data/log/make_enrollmfcc /home/iseesaw/Desktop/Share/CrossDeviceSR/kaldi/ivector-xvector-master/xvector/data/feat/mfcc
utils/validate_data_dir.sh: Successfully validated data-directory /home/iseesaw/Desktop/Share/CrossDeviceSR/kaldi/ivector-xvector-master/xvector/data
steps/make_mfcc.sh: [info]: no segments file exists: assuming wav.scp indexed by utterance.
/home/iseesaw/Desktop/Share/CrossDeviceSR/kaldi/ivector-xvector-master/xvector/utils/run.pl: line 1: parallel/run.pl: No such file or directory
Could you please tell me how to do??
Hi,
Thanks for your quick guide for ivector and xvector extraction. First of all model_3000h is not available on this link "https://pan.baidu.com/s/1kZHcOKQVkv_26emvG69IZg". So I tried to use pretrained model available here http://www.kaldi-asr.org/models/m7 but got stuck in an error.
Secondly, I am able to extract xvector on my .wav file and finally got 512-dimensional vector. Now question is that what does this 512-d vector describes ? I was supposed to get xvectors in place of MFCC frames, right ? My .wav file is of 26min duration from AMI corpus.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.