File size: 1,583 Bytes
fc4d854 71380ca fc4d854 69bd548 45b5d99 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 |
---
license: apache-2.0
datasets:
- techiaith/commonvoice_18_0_cy_en
language:
- cy
- en
base_model:
- openai/whisper-base
pipeline_tag: automatic-speech-recognition
tags:
- whisper.cpp
---
# whisper-base-ft-cv-cy-en-cpp
This model is a version of the [openai/whisper-base](https://huggingface.co./openai/whisper-base) model, fine-tuned on the
[techiaith/commonvoice_18_0_cy_en](https://huggingface.co./datasets/techiaith/commonvoice_18_0_cy_en) dataset, and then
[converted for use in whisper.cpp](https://github.com/ggerganov/whisper.cpp/tree/master/models#fine-tuned-models). Whispercpp is
a C/C++ port of Whisper that provides high performance inference on offline hardware such as desktops, laptops and mobile devices.
The model is a smaller in size to the corresponding cloud hosted model
[techiaith/whisper-large-v3-ft-cv-cy-en](https://huggingface.co./techiaith/whisper-large-v3-ft-cv-cy-en).
It achieves a success rate of 98.34% on detecting the correct language in speech,
while for transcribing it achieves the following WER results:
- Welsh: 40.10
- English: 30.9
## Usage
whispercpp makes it easy to use models in many platforms and applications. See the 'examples' folder
in the whispercpp github repo for more information and example code.
To get quickly started with whispercpp's basic usage however, follow the '[Quick Start](https://github.com/ggerganov/whisper.cpp?tab=readme-ov-file#quick-start)'
but download this model with the following command:
`$ wget https://huggingface.co./techiaith/whisper-base-ft-cv-cy-en-cpp/resolve/main/ggml-model.bin`
|