Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The Chinese result is wrong, but whisper.cpp is correct. #212

Open
j20cc opened this issue Mar 19, 2024 · 0 comments
Open

The Chinese result is wrong, but whisper.cpp is correct. #212

j20cc opened this issue Mar 19, 2024 · 0 comments

Comments

@j20cc
Copy link

j20cc commented Mar 19, 2024

use whisper.rn get

image

use whisper.cpp get

image

here is my code, they use same model and input audio

const whisper = async () => {
    const path = ReactNativeBlobUtil.fs.dirs.DocumentDir;
    const whisperContext = await initWhisper({
      filePath: `file://${path}/ggml-small.bin`,
    });

    const sampleFilePath = `file://${path}/1.wav`;
    console.log('sampleFilePath:', sampleFilePath);

    console.log('Start transcribing...');
    const startTime = Date.now();
    const {stop, promise} = whisperContext.transcribe(sampleFilePath, {
      language: 'auto',
      prompt: '以下是普通话',
      onNewSegments: segments => {
        console.log('New segments:', segments);
      },
    });
    const {result, segments} = await promise;
    console.log('result:', result);
  };

the language option i also tried 'chinese', not work correct too, whisper.cpp result is correct

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant