1 min voice data can also be used to train a good TTS model! (few shot voice cloning) Topics text-to-speechttsvoice-cloningvitsvoice-clonevoice-cloneai Resources Readme License MIT license Activity Stars 47.1ks
1 min voice data can also be used to train a good TTS model! (few shot voice cloning) - Commits · RVC-Boss/GPT-SoVITS
Permission denied in opening the [0d-语音文本校对标注工具] Try to change folder and give the authority,but still not working Platform = win10
1 min voice data can also be used to train a good TTS model! (few shot voice cloning) - RVC-Boss/GPT-SoVITS
1 min voice data can also be used to train a good TTS model! (few shot voice cloning) - Merge branch 'RVC-Boss:main' into mps · RVC-Boss/GPT-SoVITS@b942af7
"runtime\python" tools/slice_audio.py "D:\Videos\data\gpt" "output/slicer_opt" -34 4000 300 10 500 0.9 0.25 3 4 系统找不到指定的路径。 系统找不到指定的路径。 So "runtime\python" is hard coded into this project. It's not okay to just modify the go-webui.bat ...
https://github.com/RVC-Boss/GPT-SoVITS/blob/main/GPT_SoVITS/inference_webui.py#L515后面加上 del pred_semantic torch.cuda.empty_cache() 那么推理过程能否也添加这个呢。因为它在推理过程中也会积累大量的显存占用。我希望能够推理长文本也能平稳保持显存数量,而不至于会炸掉。慢点无所谓。
1 min voice data can also be used to train a good TTS model! (few shot voice cloning) - https://github.com/RVC-Boss/GPT-SoVITS/issues/1419 · Afro-ai/GPT-SoVITS@bf289e0
From what I understand, the model currently requires fine-tuning on at least 2-3 hours of speech data to produce convincing results in Japanese. Is this correct? Additionally, is it necessary to fine-tune only the SoVITS model, or does the GPT model require it as well?
我是GPT-SoVITS的忠实粉丝,经过实际使用,我认为完全可以充当有声书的配音演员。 于是乎,我就做了一个以GPT-Sovits为合成工具的有声书AI合成工具。 和微软的TTS一样,后台接受所需合成章节的SSML,形成合成任务队列。 后台依次获得任务,按照底模不同来分组合成。 一段段