Instructions to use LiquidAI/LFM3-Audio-2B-MLX-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use LiquidAI/LFM3-Audio-2B-MLX-4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download LiquidAI/LFM3-Audio-2B-MLX-4bit --local-dir LFM3-Audio-2B-MLX-4bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Download processor_config.json from LiquidAI/LFM3-Audio-2B-MLX-4bit: direct link, hf CLI and curl.
- Browser
- Download file 773 Bytes
-
https://huggingface.co/LiquidAI/LFM3-Audio-2B-MLX-4bit/resolve/main/processor_config.json
- Command line
-
hf download hf://LiquidAI/LFM3-Audio-2B-MLX-4bit/processor_config.json
-
curl -L -o processor_config.json https://huggingface.co/LiquidAI/LFM3-Audio-2B-MLX-4bit/resolve/main/processor_config.json
773 Bytes
| { | |
| "audio_out_token": "<|audio_out|>", | |
| "audio_token": "<|audio|>", | |
| "feature_extractor": { | |
| "encoder_stride": 8, | |
| "feature_extractor_type": "Lfm3AudioFeatureExtractor", | |
| "feature_size": 128, | |
| "hop_length": 160, | |
| "log_floor": -12.0, | |
| "n_fft": 400, | |
| "normalization_divisor": 4.0, | |
| "normalization_offset": 4.0, | |
| "padding_side": "right", | |
| "padding_value": 0.0, | |
| "return_attention_mask": true, | |
| "sampling_rate": 16000, | |
| "win_length": 400, | |
| "auto_map": { | |
| "AutoFeatureExtractor": "feature_extraction_lfm3_audio.Lfm3AudioFeatureExtractor" | |
| } | |
| }, | |
| "processor_class": "Lfm3AudioProcessor", | |
| "auto_map": { | |
| "AutoProcessor": "processing_lfm3_audio.Lfm3AudioProcessor" | |
| }, | |
| "audio_tokenizer_subfolder": "audio_tokenizer" | |
| } | |