Question: how does your language model work? (Inaccuracies in generated output )
parisa-hr opened this issue · comments
Hello,
I hope this message finds you well. I have been using your project to generate text from audio files.
To accomplish this, I have been following a specific process: I create an MP3 file using the website https://ttsmp3.com/, then convert the MP3 file to a WAV file.
I have seen your Document about The recognized text is wrong. but to prevent any noise or etc. I create my samples by ttsmp3.
here are my commands:
sox 1.mp3 1.wav
pocketsphinx single 1.wav > 1.json
But some of the words are incorrect like these.
I wanted to bring this issue to your attention and kindly ask for assistance or guidance on how your language model works. I am curious to understand if there are any specific steps I should follow or considerations I should keep in mind to ensure the accuracy of the generated audio output.