Comments (10)
jcsilva
commented on July 17, 2024
Did you change the paths configure in models/conf/ivector_extractor.conf file?
from docker-kaldi-gstreamer-server.
cbsandeep10
commented on July 17, 2024
I checked the path of the files in ivector twice.
Everything seems to be fine.
from docker-kaldi-gstreamer-server.
jcsilva
commented on July 17, 2024
Did you download this model from somewhere? I look for it here (http://kaldi-asr.org/downloads/tree/trunk/egs/librispeech/s5/exp/) and here (https://api.ai/downloads/api.ai-kaldi-asr-model.zip), but I couldn't donwload it.
I would like to test here and find out the problem...
from docker-kaldi-gstreamer-server.
cbsandeep10
commented on July 17, 2024
When I said Librispeech, I meant that I trained the model with our custom data using the below script
https://github.com/kaldi-asr/kaldi/blob/master/egs/librispeech/s5/local/chain/run_tdnn.sh.
I didn't download the official kaldi models.
Is there any other parameters that I have to change related to this chain tdnn model.
from docker-kaldi-gstreamer-server.
cbsandeep10
commented on July 17, 2024
I guess I need to use a different gst plugin for nnet3 chain models.
I am currently using this plugin /opt/gst-kaldi-nnet2-online/src/libgstkaldionline2.so.
from docker-kaldi-gstreamer-server.
jcsilva
commented on July 17, 2024
I've just tried here with a chain model and the docker image from docker hub. As you can see below, everything was fine. This chain model was trained in September 2017. I don't know when you trained your model and I don't know if there was any change in Kaldi file's format recently. Maybe, if there was such a change, the docker hub image can be outdated. So, you could try to build your own docker image. I don't know if it will solve your problem... it is just a guess.
It is my yaml file:
use-nnet2: True
decoder:
# All the properties nested here correspond to the kaldinnet2onlinedecoder GStreamer plugin properties.
# Use gst-inspect-1.0 ./libgstkaldionline2.so kaldinnet2onlinedecoder to discover the available properties
nnet-mode: 3
use-threaded-decoder: True
model : /opt/models/tdnn1c_rvb_lstm_sp_online/final.mdl
word-syms : /opt/models/words.txt
fst : /opt/models/HCLG.fst
mfcc-config : /opt/models/tdnn1c_rvb_lstm_sp_online/conf/mfcc.conf
ivector-extraction-config : /opt/models/tdnn1c_rvb_lstm_sp_online/conf/ivector_extractor.conf
frame-subsampling-factor: 3
#frames-per-chunk: 20
#extra-left-context-initial: 0
max-active: 10000
beam: 15.0
lattice-beam: 8.0
acoustic-scale: 1.25
# lmwt-scale: 0.8
do-endpointing : false
##endpoint-silence-phones : "1:2:3:4:5:6:7:8:9:10"
traceback-period-in-secs: 0.25
chunk-length-in-secs: 0.25
num-nbest: 10
#Additional functionality that you can play with:
# lm-fst: /opt/models/tdnn1c_rvb_lstm_sp_online/lang_small9_carpa/G.fst
# big-lm-const-arpa: /opt/models/tdnn1c_rvb_lstm_sp_online/lang_small9_carpa/G.carpa
# phone-syms: /opt/models/tdnn1c_rvb_lstm_sp_online/lang_small9_carpa/phones.txt
# word-boundary-file: /opt/models/tdnn1c_rvb_lstm_sp_online/lang_small9_carpa/phones/word_boundary.int
# do-phone-alignment: true
out-dir: /opt/tmp
use-vad: False
silence-timeout: 10
# Just a sample post-processor that appends "." to the hypothesis
post-processor: perl -npe 'BEGIN {use IO::Handle; STDOUT->autoflush(1);} s/(.*)/\1/;'
# A sample full post processor that add a confidence score to 1-best hyp and deletes other n-best hyps
#full-post-processor: /opt/kaldi-gstreamer-server/sample_full_post_processor.py
logging:
version : 1
disable_existing_loggers: False
formatters:
simpleFormater:
format: '%(asctime)s - %(levelname)7s: %(name)10s: %(message)s'
datefmt: '%Y-%m-%d %H:%M:%S'
handlers:
console:
class: logging.StreamHandler
formatter: simpleFormater
level: DEBUG
root:
level: DEBUG
handlers: [console]
and this was the output (worker.log):
root@a0750856fced:/opt# cat worker.log
libdc1394 error: Failed to initialize libdc1394
libudev: udev_has_devtmpfs: name_to_handle_at on /dev: Operation not permitted
DEBUG 2018-06-12 13:04:55,463 Starting up worker
2018-06-12 13:04:55 - INFO: decoder2: Creating decoder using conf: {'post-processor': "perl -npe 'BEGIN {use IO::Handle; STDOUT->autoflush(1);} s/(.*)/\\1/;'", 'logging': {'version': 1, 'root': {'level': 'DEBUG', 'handlers': ['console']}, 'formatters': {'simpleFormater': {'datefmt': '%Y-%m-%d %H:%M:%S', 'format': '%(asctime)s - %(levelname)7s: %(name)10s: %(message)s'}}, 'disable_existing_loggers': False, 'handlers': {'console': {'formatter': 'simpleFormater', 'class': 'logging.StreamHandler', 'level': 'DEBUG'}}}, 'use-vad': False, 'decoder': {'ivector-extraction-config': '/opt/models/tdnn1c_rvb_lstm_sp_online/conf/ivector_extractor.conf', 'lattice-beam': 8.0, 'acoustic-scale': 1.25, 'do-endpointing': False, 'beam': 15.0, 'mfcc-config': '/opt/models/tdnn1c_rvb_lstm_sp_online/conf/mfcc.conf', 'traceback-period-in-secs': 0.25, 'nnet-mode': 3, 'word-syms': '/opt/models/words.txt', 'num-nbest': 10, 'frame-subsampling-factor': 3, 'max-active': 10000, 'fst': '/opt/models/HCLG.fst', 'use-threaded-decoder': True, 'model': '/opt/models/tdnn1c_rvb_lstm_sp_online/final.mdl', 'chunk-length-in-secs': 0.25}, 'silence-timeout': 10, 'out-dir': '/opt/tmp', 'use-nnet2': True}
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: nnet-mode = 3
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: ivector-extraction-config = /opt/models/tdnn1c_rvb_lstm_sp_online/conf/ivector_extractor.conf
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: lattice-beam = 8.0
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: acoustic-scale = 1.25
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: do-endpointing = False
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: beam = 15.0
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: mfcc-config = /opt/models/tdnn1c_rvb_lstm_sp_online/conf/mfcc.conf
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: traceback-period-in-secs = 0.25
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: word-syms = /opt/models/words.txt
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: num-nbest = 10
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: frame-subsampling-factor = 3
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: max-active = 10000
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: fst = /opt/models/HCLG.fst
2018-06-12 13:04:55 - INFO: decoder2: Setting decoder property: model = /opt/models/tdnn1c_rvb_lstm_sp_online/final.mdl
LOG ([5.2.187~1-148c8]:CompileLooped():nnet-compile-looped.cc:336) Spent 0.016602 seconds in looped compilation.
2018-06-12 13:04:56 - INFO: decoder2: Setting decoder property: chunk-length-in-secs = 0.25
2018-06-12 13:04:56 - INFO: decoder2: Created GStreamer elements
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstAppSrc object at 0x7f2ac23ec3c0 (GstAppSrc at 0x21727b0)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstDecodeBin object at 0x7f2ac23ec2d0 (GstDecodeBin at 0x21e4060)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstAudioConvert object at 0x7f2ac23ec280 (GstAudioConvert at 0x21ee4e0)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstAudioResample object at 0x7f2ac23ec370 (GstAudioResample at 0x1fb7430)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstTee object at 0x7f2ac23ec230 (GstTee at 0x21fc000)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstQueue object at 0x7f2ac23ec320 (GstQueue at 0x2200170)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstFileSink object at 0x7f2ac23ec410 (GstFileSink at 0x2203200)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstQueue object at 0x7f2ac23ec460 (GstQueue at 0x2200460)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.Gstkaldinnet2onlinedecoder object at 0x7f2ac23ec4b0 (Gstkaldinnet2onlinedecoder at 0x221e150)> to the pipeline
2018-06-12 13:04:56 - DEBUG: decoder2: Adding <__main__.GstFakeSink object at 0x7f2ac23ec500 (GstFakeSink at 0x2038440)> to the pipeline
2018-06-12 13:04:56 - INFO: decoder2: Linking GStreamer elements
LOG ([5.2.187~1-148c8]:ComputeDerivedVars():ivector-extractor.cc:183) Computing derived variables for iVector extractor
LOG ([5.2.187~1-148c8]:ComputeDerivedVars():ivector-extractor.cc:204) Done.
2018-06-12 13:04:56 - INFO: decoder2: Setting pipeline to READY
2018-06-12 13:04:56 - INFO: decoder2: Set pipeline to READY
2018-06-12 13:04:56 - INFO: __main__: Opening websocket connection to master server
2018-06-12 13:04:56 - INFO: __main__: Opened websocket connection to server
from docker-kaldi-gstreamer-server.
cbsandeep10
commented on July 17, 2024
Mine is a very recent one. I trained the model in April 2018.
from docker-kaldi-gstreamer-server.
jcsilva
commented on July 17, 2024
I'm also using /opt/gst-kaldi-nnet2-online/src/libgstkaldionline2.so plugin. I don't believe you need to change it.
Some suggestions:
- Be sure you can run a decoding with you model in your host machine. You should use a tool such as online2-wav-nnet3...
- If step 1 is ok, try to build your own docker image. The versioned Dockerfile will download the most recent kaldi version and will compile it.
- Test your model with this brand new image.
I think this way you have reasonable chances to solve your problem.
from docker-kaldi-gstreamer-server.
cbsandeep10
commented on July 17, 2024
Thanks @jcsilva for helping me out.
- Yes,I can decode with online2-wav-nnet3-latgen-faster with the chain model locally.
- I will build new docker image from the dockerfile and let you know.
from docker-kaldi-gstreamer-server.
cbsandeep10
commented on July 17, 2024
Thank you so much @jcsilva .
I built a new docker image and It worked perfectly.
from docker-kaldi-gstreamer-server.
Related Issues (20)
- Can I update the Kaldi version in the Docker container? HOT 2
- Unknown component type TdnnComponent HOT 2
- response['transcript'] did not return final:true
- Build fails: MKL libraries could not be found HOT 1
- I want to get the start and end time of the sentence.
- No decoder available ,try again later HOT 1
- greek characters
- Cannot build on a Raspberry Pi 4 HOT 1
- recognition occurs, no result returned to client, then worker/master hangs HOT 2
- Portuguese model avaliable
- How to api provided with ajax request
- core happened using chain model HOT 3
- Command in README.md causes incorrect rocognition on speech containing "test"
- How to tune Kaldi gsteamer for using NNET3 decoder
- ReadFstKaldiGeneric():kaldi-fst-io.cc:82) Could not read fst from {MODEL_FILE}
- Docker build errors on Ubuntu 18.04 HOT 1
- No responses returned on websocket
- worker is errors
- Received error from server (status 9) Error message: No decoder available, try again later HOT 1
- failed: Connection refused. wget https://phon.ioc.ee/~tanela/tedlium_nnet_ms_sp_online.tgz** HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from docker-kaldi-gstreamer-server.