Instructions to use gafiatulin/tada-codec-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use gafiatulin/tada-codec-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir tada-codec-mlx gafiatulin/tada-codec-mlx
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
metadata
license: mit
base_model: HumeAI/tada-codec
tags:
- mlx
- audio-codec
- tada
TADA Codec MLX
MLX-converted weights for HumeAI/tada-codec. Part of tada-mlx.
Contains the shared decoder, encoder, and aligners used by both tada-1b-mlx and tada-3b-ml-mlx.
Contents
decoder.safetensors— DAC waveform decoder with local attentionencoder.safetensors— WavEncoder + local attention encoderaligner.safetensors— wav2vec2-large CTC aligner (English)aligner-*.safetensors— Multilingual aligners (ar, ch, de, es, fr, it, ja, pl, pt)