Downloads · 30 days
0
thonburian-whisper-ggml
suebphatt/thonburian-whisper-ggml
thonburian-whisper-ggml is a automatic speech recognition model from suebphatt. Use it when you need speech turned into text. The card lists the license as apache-2.0.
Pre-converted, Q50-quantized GGML files of Thonburian Whisper (Thai fine-tuned Whisper by BiodatLab) for use with whisper.cpp.
Access
Public
Updated Sep 18, 2026
Repo size
1.6 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin1.6 GB · 100%
At a glance
- Task
- Automatic Speech Recognition
- License
- apache-2.0
- Access
- Public
- Created
- Sep 18, 2026
- Updated
- Sep 18, 2026
- SHA
- d271c270
Base models
Model card
From the Hugging Face model README
Thonburian Whisper — GGML (whisper.cpp) builds
Pre-converted, Q5_0-quantized GGML files of Thonburian Whisper (Thai fine-tuned Whisper by BiodatLab) for use with whisper.cpp.
| File | Source model | Size |
|---|---|---|
thonburian-large-v3-q5_0.bin | biodatlab/whisper-th-large-v3-combined | 1.03 GB |
thonburian-medium-q5_0.bin | biodatlab/whisper-th-medium-combined | 539 MB |
Conversion: models/convert-h5-to-ggml.py from whisper.cpp (F16), then whisper-quantize ... q5_0.
No weights were changed beyond quantization. SHA-256 checksums are in checksums.sha256.
Usage:
whisper-cli -m thonburian-medium-q5_0.bin -l th -f audio.wav
Built for OnTheKey (offline Thai dictation). License follows the upstream models: Apache-2.0.
- Task
- Automatic Speech Recognition
- License
- apache-2.0
- Languages
- th
- Created
- Sep 18, 2026
- Updated
- Sep 18, 2026