ggml-whisper-unofficial: Curated Fine-Tuned Whisper Models
A continuously updated repository collecting non-OpenAI official fine-tuned Whisper models converted to ggml
format (used in whisper.cpp).
All models are converted from community fine-tuned versions, focusing on domain-specific performance (e.g., anime voice, industry-specific speech) and compatibility with CPU/GPU inference via whisper.cpp. New models will be added regularly.
Model in this repository
| Model | Ref (Original Fine-Tuned Model) | SHA-256 | Size |
|---|---|---|---|
| ggml-bagasshw-whisper-large-v2.bin | bagasshw/whisper-large-v2 | e954f27623d162b81dd02542fbf5658fc9d7627db02484799779066e43127556 |
3.09 GB |
| ggml-litagin-anime-whisper.bin | litagin/anime-whisper | bce42c15312fcdbdcc8432b2ce63ba967f4cc3d2cc783fd6ab2ba264caf0db8d |
1.52 GB |
| ggml-whisper-hindi-large-v2.bin | vasista22/whisper-hindi-large-v2 | 62cc7c4135faad16ac412c2f2e2e9f761a94c2b64b4e8b6a88c1ef2b8b5837ae |
3.09 GB |
| ggml-whisper-large-v2-ru.bin | mitchelldehaven/whisper-large-v2-ru | cf16dfa83c23c6697379b676223c2a03df27ca549b536bd9434722350c72a8f0 |
3.09 GB |
| ggml-whisper-large-v2-uk.bin | mitchelldehaven/whisper-large-v2-uk | 0e351d0c45dddd42b079c0a1498d4b72fde442565a71fbe5872a49782acb96d7 |
3.09 GB |
| ggml-whisper-tamil-large-v2.bin | vasista22/whisper-tamil-large-v2 | 80ccc7c0fbed733e7df98ff9bb4d7d6316e461fc3096283e21c3514f9f00ca4b |
3.09 GB |
| ggml-whisper-telugu-large-v2.bin | vasista22/whisper-telugu-large-v2 | 0bdcf1e093c9802c2a873a87b7d6db21f537fcd72a6394ff68e103299b70123c |
3.09 GB |
Model from other repositories
| repository | URL | License |
|---|---|---|
| kotoba-tech/kotoba-whisper-v1.0-ggml | https://huggingface.co/kotoba-tech/kotoba-whisper-v1.0-ggml | apache-2.0 |
| kotoba-tech/kotoba-whisper-v2.0-ggml | https://huggingface.co/kotoba-tech/kotoba-whisper-v2.0-ggml | apache-2.0 |
| distil-whisper/distil-large-v3-ggml | https://huggingface.co/distil-whisper/distil-large-v3-ggml | mit |
| distil-whisper/distil-large-v3.5-ggml | https://huggingface.co/distil-whisper/distil-large-v3.5-ggml | mit |
| distil-whisper/distil-large-v2 | https://huggingface.co/distil-whisper/distil-large-v2 | mit |
| distil-whisper/distil-small.en | https://huggingface.co/distil-whisper/distil-small.en | mit |
| distil-whisper/distil-medium.en | https://huggingface.co/distil-whisper/distil-medium.en | mit |
| xezpeleta/whisper-large-v3-eu | https://huggingface.co/xezpeleta/whisper-large-v3-eu | apache-2.0 |
Note: You can clone the whisper PyTorch models from the original repositories and convert them to
ggmlformat using the whisper.cpp | convert-h5-to-ggml.py
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support