| license: mit | |
| library_name: pytorch | |
| tags: | |
| - audio | |
| - speech | |
| - neural-audio-codec | |
| - scale-invariance | |
| - shape-gain-decomposition | |
| - encodec | |
| - bigcodec | |
| pipeline_tag: audio-to-audio | |
| # The Equalizer: Checkpoints | |
| Pretrained model checkpoints for **The Equalizer: Shape-Gain Decomposition in Neural Audio Codecs**. | |
| For implementation and usage instructions, please refer to the GitHub repository: | |
| 👉 [GitHub Repository](https://github.com/samir-sadok/Equalizer) (or your public GitHub link) | |
| For details, refer to the paper: | |
| > [The Equalizer: Introducing Shape-Gain Decomposition in Neural Audio Codecs](https://arxiv.org/pdf/2602.15491) | |
| ## Checkpoints Directory Structure | |
| This repository contains the following checkpoints: | |
| - `EnCodec/1024/base/`: Standard base EnCodec model (SpeechTokenizer). | |
| - `EnCodec/1024/eq/`: EnCodec model fine-tuned using The Equalizer (Shape-Gain Decomposition). | |
| - `BigCodec/8192/base/`: Standard base BigCodec model. | |
| - `BigCodec/8192/eq/`: BigCodec model fine-tuned using The Equalizer (Shape-Gain Decomposition). | |