Choosing a Whisper Model
Myna supports multiple Whisper model sizes. Larger models produce more accurate transcriptions but require more disk space and take longer to process.
Available Models
| Model | Size | Speed | Accuracy | Best For |
|---|---|---|---|---|
| base | 150 MB | Fastest | Good | Clear audio, quick results, lower-spec hardware |
| small | 488 MB | Medium | Better | Accented speech, moderate background noise |
| large-v3-turbo (Recommended) | 1.6 GB | Medium | Excellent | Best overall balance of speed and accuracy |
| large-v3 | 3.1 GB | Slowest | Maximum | Difficult audio, maximum accuracy needed |
Downloading Models
- Go to Settings > Transcription.
- Select the model you want to download.
- Click Download. The download will run in the background.
The base model is included with the installer. All other models are optional downloads.
Custom Model Storage Path
By default, models are stored in Myna's application data folder. You can change the storage location in Settings > Transcription > Model Storage Path if you need to save space on your main drive.
Comments
0 comments