| kokoro-tts input.txt output.wav --speed 1.2 --lang en-us --voice af_sarah | |
| https://github.com/hexgrad/kokoro | |
| 看这里 各种语言都有 | |
| https://huggingface.co/hexgrad/Kokoro-82M | |
| ja 不正常，从这里下载模型试试 | |
| https://huggingface.co/spaces/hexgrad/Kokoro-TTS | |
| 这里找运行代码 | |
| https://github.com/hexgrad/misaki | |
| https://huggingface.co/spaces/NeuralFalcon/KOKORO-TTS-1.0 | |
| 看它生成的 ja audio 怎么样 | |
| https://github.com/nazdridoy/kokoro-tts/issues/1 | |
| 为什么用不了 gpu | |
| cuda_12.0.0_527.41_windows.exe 安装 | |
| pip install nvidia-cudnn-cu12 | |
| conda install -y cudnn 再安装 | |
| https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements | |
| git clone https://github.com/nazdridoy/kokoro-tts.git \ | |
| && cd kokoro-tts \ | |
| && conda create -n kokorotts python=3.10 pip \ | |
| && conda activate kokorotts \ | |
| && pip install -r requirements.txt | |
| wget https://github.com/nazdridoy/kokoro-tts/releases/download/v1.0.0/voices-v1.0.bin | |
| wget https://github.com/nazdridoy/kokoro-tts/releases/download/v1.0.0/kokoro-v1.0.onnx | |
| 两个文件放根目录 | |
| ``` | |
| | **Category** | **Voices** | **Language Code** | | |
| | ------------ | ------------------------------------------------------------ | ----------------- | | |
| | 🇺🇸 👩 | af_alloy, af_aoede, af_bella, af_heart, af_jessica, af_kore, af_nicole, af_nova, af_river, af_sarah, af_sky | **en-us** | | |
| | 🇺🇸 👨 | am_adam, am_echo, am_eric, am_fenrir, am_liam, am_michael, am_onyx, am_puck | **en-us** | | |
| | 🇬🇧 | bf_alice, bf_emma, bf_isabella, bf_lily, bm_daniel, bm_fable, bm_george, bm_lewis | **en-gb** | | |
| | 🇫🇷 | ff_siwis | **fr-fr** | | |
| | 🇮🇹 | if_sara, im_nicola | **it** | | |
| | 🇯🇵 | jf_alpha, jf_gongitsune, jf_nezumi, jf_tebukuro, jm_kumo | **ja** | | |
| | 🇨🇳 | zf_xiaobei, zf_xiaoni, zf_xiaoxiao, zf_xiaoyi, zm_yunjian, zm_yunxi, zm_yunxia, zm_yunyang | **cmn** | | |
| ``` |