Dojo Speech v1.0 (ggml)
On-device speech recognition models used by Dojo Workspace for dictation and local transcription. Everything runs privately on your computer.
These are OpenAI Whisper models in the ggml format used by whisper.cpp. The files are byte-for-byte identical to the upstream whisper.cpp conversions; only the file names differ.
| File | Upstream file | Language | Size (bytes) | SHA-256 |
|---|---|---|---|---|
Dojo-Speech-v1.0-Tiny-EN.bin |
ggml-tiny.en.bin |
English | 77704715 | 921e4cf8686fdd993dcd081a5da5b6c365bfde1162e72b08d75ac75289920b1f |
Dojo-Speech-v1.0-Base-EN.bin |
ggml-base.en.bin |
English | 147964211 | a03779c86df3323075f5e796cb2ce5029f00ec8869eee3fdfb897afe36c6d002 |
Dojo-Speech-v1.0-Small-EN.bin |
ggml-small.en.bin |
English | 487614201 | c6138d6d58ecc8322097e0f987c32f1be8bb0a18532a3f88f734d1bbf9c41e5d |
Dojo-Speech-v1.0-Medium-EN.bin |
ggml-medium.en.bin |
English | 1533774781 | cc37e93478338ec7700281a7ac30a10128929eb8f427dda2e865faa8f6da4356 |
Dojo-Speech-v1.0-Large-v3-Turbo.bin |
ggml-large-v3-turbo.bin |
Multilingual | 1624555275 | 1fc70f774d38eb169993ac391eea357ef47c88757ef72ee5943879b7e8e2bc69 |
Dojo pins this repository to a specific commit and checks every file against its SHA-256 before use.
Credits and license
- Whisper models by OpenAI: https://github.com/openai/whisper, released under the MIT License.
- ggml conversions by the whisper.cpp project: https://github.com/ggml-org/whisper.cpp, released under the MIT License.
This repository is distributed under the MIT License. See LICENSE.
Model tree for DojoWorkspace/Dojo-Speech-v1.0-ggml
Base model
openai/whisper-base.en