| --- |
| license: cc-by-nc-4.0 |
| language: |
| - ja |
| base_model: |
| - akkikiki/j-moshi-ext-mlx |
| base_model_relation: quantized |
| library_name: moshi |
| tags: |
| - 8-bit precision |
| --- |
| |
| The model in this repository is an MLX model converted from the pre-trained J-Moshi model for Mac OS. |
| For other terms and restrictions, it adheres to the original models `nu-dialogue/j-moshi-ext` and `kyutai/moshi`. |
|
|
| このリポジトリのモデルは，J-Moshiの学習済みモデルをMac OS向けに変換したMLXモデルです。 |
| その他規約・制限等においてはオリジナルのモデルである`nu-dialogue/j-moshi-ext`並びに`kyutai/moshi`に準拠します。 |
|
|
| ### Installation |
| ``` |
| pip install moshi_mlx |
| ``` |
| 詳細はhttps://github.com/kyutai-labs/moshi?tab=readme-ov-file#requirements を参照。 |
|
|
| ### Example Usage |
| ``` |
| python -m moshi_mlx.local --hf-repo akkikiki/j-moshi-ext-mlx-q8 --quantized 8 |
| ``` |
