Automatic Speech Recognition
MLX
meralion3
rotorquant
kv-cache-quantization
meralion
speech-to-text
multimodal
audio
quantized
2bit
apple-silicon
custom_code
Instructions to use majentik/MERaLiON-3-10B-RotorQuant-MLX-2bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use majentik/MERaLiON-3-10B-RotorQuant-MLX-2bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download majentik/MERaLiON-3-10B-RotorQuant-MLX-2bit --local-dir MERaLiON-3-10B-RotorQuant-MLX-2bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Download preprocessor_config.json from majentik/MERaLiON-3-10B-RotorQuant-MLX-2bit: direct link, hf CLI and curl.
- Browser
- Download file 443 Bytes
-
https://huggingface.co/majentik/MERaLiON-3-10B-RotorQuant-MLX-2bit/resolve/main/preprocessor_config.json
- Command line
-
hf download hf://majentik/MERaLiON-3-10B-RotorQuant-MLX-2bit/preprocessor_config.json
-
curl -L -o preprocessor_config.json https://huggingface.co/majentik/MERaLiON-3-10B-RotorQuant-MLX-2bit/resolve/main/preprocessor_config.json
443 Bytes
| { | |
| "auto_map": { | |
| "AutoProcessor": "processing_meralion3.MERaLiON3Processor" | |
| }, | |
| "chunk_length": 30, | |
| "dither": 0.0, | |
| "feature_extractor_type": "WhisperFeatureExtractor", | |
| "feature_size": 128, | |
| "hop_length": 160, | |
| "n_fft": 400, | |
| "n_samples": 480000, | |
| "nb_max_frames": 3000, | |
| "padding_side": "right", | |
| "padding_value": 0.0, | |
| "processor_class": "MERaLiON3Processor", | |
| "return_attention_mask": false, | |
| "sampling_rate": 16000 | |
| } | |