| language: | |
| - fi | |
| license: apache-2.0 | |
| library_name: mlx | |
| tags: | |
| - automatic-speech-recognition | |
| - whisper | |
| - mlx | |
| - 4-bit | |
| - fi | |
| base_model: Finnish-NLP/whisper-medium-finnish | |
| pipeline_tag: automatic-speech-recognition | |
| # whisper-medium-finnish (MLX, 4-bit quantized) | |
| MLX conversion of [Finnish-NLP/whisper-medium-finnish](https://huggingface.co/Finnish-NLP/whisper-medium-finnish) for fast inference on Apple Silicon via [mlx-whisper](https://github.com/ml-explore/mlx-examples/tree/main/whisper). | |
| Converted with `mlx-examples/whisper/convert.py` at `4-bit quantized` precision. | |
| ## Usage | |
| ```bash | |
| pip install mlx-whisper | |
| mlx_whisper audio_file.mp3 --model FredrikKarlssonSpeech/whisper-medium-finnish-mlx-4bit | |
| ``` | |
| ```python | |
| import mlx_whisper | |
| result = mlx_whisper.transcribe("audio_file.mp3", path_or_hf_repo="FredrikKarlssonSpeech/whisper-medium-finnish-mlx-4bit") | |
| print(result["text"]) | |
| ``` | |