Vosk model vosk-model-pt-fb-v0.1.1-20220516_2113 throwing error: ConstArpaLm <LmStates> section reading failed.

[heiytor]

I am currently encountering an issue with the Vosk model vosk-model-pt-fb-v0.1.1-20220516_2113, which can be found at this link.

While attempting to use this model, I receive the following error:

ERROR (VoskAPI:ReadInternal():const-arpa-lm.cc:610) ConstArpaLm <LmStates
> section reading failed.

To address this issue, I referred to Issue #1028 and tried reinstalling the model, which did not resolve the error. I also tried cloning directly from the source, but this also resulted in the same error.

Below is the full output:

LOG (VoskAPI:ReadDataFiles():model.cc:213) Decoding params beam=10 max-ac
tive=7000 lattice-beam=6
LOG (VoskAPI:ReadDataFiles():model.cc:216) Silence phones 1:2:3:4:5:6:7:8
:9:10
LOG (VoskAPI:RemoveOrphanNodes():nnet-nnet.cc:948) Removed 0 orphan nodes
.
LOG (VoskAPI:RemoveOrphanComponents():nnet-nnet.cc:847) Removing 0 orphan
 components.
LOG (VoskAPI:ReadDataFiles():model.cc:248) Loading i-vector extractor fro
m D:\presicion-data\vosk-api\vosk-model-pt-fb-v0.1.1-20220516_2113/ivecto
r/final.ie
LOG (VoskAPI:ComputeDerivedVars():ivector-extractor.cc:183) Computing der
ived variables for iVector extractor
LOG (VoskAPI:ComputeDerivedVars():ivector-extractor.cc:204) Done.        
LOG (VoskAPI:ReadDataFiles():model.cc:279) Loading HCLG from D:\presicion
-data\vosk-api\vosk-model-pt-fb-v0.1.1-20220516_2113/graph/HCLG.fst      
LOG (VoskAPI:ReadDataFiles():model.cc:294) Loading words from D:\presicio
n-data\vosk-api\vosk-model-pt-fb-v0.1.1-20220516_2113/graph/words.txt    
LOG (VoskAPI:ReadDataFiles():model.cc:303) Loading winfo D:\presicion-dat
a\vosk-api\vosk-model-pt-fb-v0.1.1-20220516_2113/graph/phones/word_bounda
ry.int
LOG (VoskAPI:ReadDataFiles():model.cc:310) Loading subtract G.fst model f
rom D:\presicion-data\vosk-api\vosk-model-pt-fb-v0.1.1-20220516_2113/resc
ore/G.fst
LOG (VoskAPI:ReadDataFiles():model.cc:312) Loading CARPA model from D:\pr
esicion-data\vosk-api\vosk-model-pt-fb-v0.1.1-20220516_2113/rescore/G.car
pa
ERROR (VoskAPI:ReadInternal():const-arpa-lm.cc:610) ConstArpaLm <LmStates
> section reading failed.

Additionally, I tested the vosk-model-pt-fb-v0.1.1-20220516_2113 model available here. However, the output indicated that my microphone was not recognized, as shown below:

$ node index.js 
LOG (VoskAPI:ReadDataFiles():model.cc:213) Decoding params beam=13 max-ac
tive=7000 lattice-beam=6
LOG (VoskAPI:ReadDataFiles():model.cc:216) Silence phones 1:2:3:4:5:6:7:8
:9:10
LOG (VoskAPI:RemoveOrphanNodes():nnet-nnet.cc:948) Removed 1 orphan nodes
.
LOG (VoskAPI:RemoveOrphanComponents():nnet-nnet.cc:847) Removing 2 orphan
 components.
LOG (VoskAPI:Collapse():nnet-utils.cc:1488) Added 1 components, removed 2
LOG (VoskAPI:ReadDataFiles():model.cc:248) Loading i-vector extractor fro
m D:\presicion-data\vosk-api\vosk-model-small-pt-0.3/ivector/final.ie    
LOG (VoskAPI:ComputeDerivedVars():ivector-extractor.cc:183) Computing der
ived variables for iVector extractor
LOG (VoskAPI:ComputeDerivedVars():ivector-extractor.cc:204) Done.        
LOG (VoskAPI:ReadDataFiles():model.cc:282) Loading HCL and G from D:\pres
icion-data\vosk-api\vosk-model-small-pt-0.3/HCLr.fst D:\presicion-data\vo
sk-api\vosk-model-small-pt-0.3/Gr.fst
LOG (VoskAPI:ReadDataFiles():model.cc:303) Loading winfo D:\presicion-dat
a\vosk-api\vosk-model-small-pt-0.3/word_boundary.int
{ partial: '' }
{ partial: '' }
{ partial: '' }
{ partial: '' }
{ partial: '' }
{ text: '' }
{ partial: '' }
{ partial: '' }
{ partial: '' }
{ partial: '' }
{ text: '' }
{ partial: '' }

Stopping

I am currently using Windows 11, Sox 14.2.2, and Node.js 18.16.0. Below is the full code:

const path = require('path')
var vosk = require('vosk')
const fs = require("fs");
var mic = require("mic");

const MODEL_1 = path.resolve(__dirname, 'vosk-model-pt-fb-v0.1.1-20220516_2113');
const MODEL_2 = path.resolve(__dirname, 'vosk-model-small-pt-0.3')

SAMPLE_RATE = 16000

if (!fs.existsSync(MODEL_1)) {
    process.exit()
}

vosk.setLogLevel(0);
const model = new vosk.Model(MODEL_1);
const rec = new vosk.Recognizer({model: model, sampleRate: SAMPLE_RATE, frameSize: 4000});

var micInstance = mic({
    rate: String(SAMPLE_RATE),
    channels: '1',
    debug: false,
    device: 'default',    
});

var micInputStream = micInstance.getAudioStream();

micInputStream.on('data', data => {
    if (rec.acceptWaveform(data)) {
        console.log(rec.result());
    } else{
        console.log(rec.partialResult());
    }
});

micInputStream.on('audioProcessExitComplete', function() {
    console.log("Cleaning up");
    console.log(rec.finalResult());
    rec.free();
    model.free();
});

process.on('SIGINT', function() {
    console.log("\nStopping");
    micInstance.stop();
});

micInstance.start();

Any assistance in addressing this issue would be greatly appreciated.

Thank you.

[nshmyrev]

You need pruned version as described in

https://alphacephei.com/vosk/models/vosk-model-pt-fb-v0.1.1-pruned.zip

#991 (comment)

essentially it is the same issue as #991

As for microphone issue, you'd better use node-portaudio module, not mic module, it should be more stable. We have issue #841

[brunoreisportela]

I was a quite while looking for this file. Thanks for sharing @nshmyrev