Whisper Models
VoicePad uses faster-whisper, a CTranslate2-optimised implementation of OpenAI Whisper.
Default Model
The default model is turbo, a fine-tuned version of large-v3 that is significantly faster while maintaining high accuracy. It is the recommended model for most users with an NVIDIA GPU.
Models Shown in the UI
VoicePad now shows a short curated model list in onboarding and the Settings tab:
| UI Label | Model ID | Best For |
|---|---|---|
| Recommended - Turbo | turbo |
Most NVIDIA GPU users |
| Faster - Small | small |
Lower-VRAM GPUs and lighter workloads |
| Highest Accuracy - Large v3 | large-v3 |
Users who want maximum accuracy and have more VRAM |
This keeps the interface simple for most people while still exposing the best common choices.
Model Comparison
| Model | Parameters | VRAM | Speed | Accuracy | Language |
|---|---|---|---|---|---|
tiny |
39 M | < 1 GB | Very fast | Low | Multilingual |
base |
74 M | < 1 GB | Fast | Fair | Multilingual |
small |
244 M | ~1 GB | Moderate | Good | Multilingual |
medium |
769 M | ~2 GB | Slow | Very good | Multilingual |
large-v2 |
1.5 B | ~5 GB | Very slow | Excellent | Multilingual |
large-v3 |
1.5 B | ~5 GB | Very slow | Excellent | Multilingual |
turbo |
809 M | ~3 GB | Moderate | Excellent | Multilingual |
distil-small.en |
134 M | < 1 GB | Fast | Good | English only |
distil-medium.en |
394 M | ~1 GB | Moderate | Very good | English only |
distil-large-v2 |
756 M | ~3 GB | Slow | Excellent | English only |
distil-large-v3 |
756 M | ~3 GB | Slow | Excellent | English only |
Which Model Should I Use?
| Your Hardware | Recommended Model |
|---|---|
| NVIDIA GPU, 4 GB VRAM | turbo |
| NVIDIA GPU, 6-8 GB VRAM | turbo or large-v3 |
| NVIDIA GPU, < 4 GB VRAM | small or distil-small.en |
| CPU only | small or base |
Advanced Models
VoicePad core supports a broader set of compatible Whisper-family models than the simple UI list shows. If you want to use an advanced model, edit ~/.config/voicepad/voicepad.yaml directly and set transcription_model to the model ID you want.
This is useful for power users who want to try smaller, English-only, distil, or other compatible CTranslate2-converted checkpoints without adding complexity to the default interface.
Changing the Model
For the common path, open the Settings tab in VoicePad, select one of the curated models from the dropdown, and press Save. VoicePad reloads the model immediately.
For advanced models, edit voicepad.yaml directly and restart VoicePad if needed.
The first time you use a new model, it downloads from HuggingFace. Download sizes range from ~75 MB (tiny) to ~3 GB (large-v3). After the first download, the model is cached locally and no network access is required.