blackteay/GPT-SoVITS
1
0
Fork 0
mirror of https://github.com/RVC-Boss/GPT-SoVITS.git synced 2025-04-06 03:57:44 +08:00
Code Issues Packages Projects Releases Wiki Activity

support speed adjustment

Browse Source 
support speed adjustment
This commit is contained in:
RVC-Boss 2024-07-22 18:23:33 +08:00 committed by GitHub
parent 4f8e1660af
commit f4f76ea431
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
GPT_SoVITS/inference_webui.py
View File

@ -312,7 +312,7 @@ def merge_short_text_in_array(texts, threshold):
result[len(result) - 1] += text result[len(result) - 1] += text
return result return result
def get_tts_wav(ref_wav_path, prompt_text, prompt_language, text, text_language, how_to_cut=i18n("不切"), top_k=20, top_p=0.6, temperature=0.6, ref_free = False): def get_tts_wav(ref_wav_path, prompt_text, prompt_language, text, text_language, how_to_cut=i18n("不切"), top_k=20, top_p=0.6, temperature=0.6, ref_free = False,speed=1):
if prompt_text is None or len(prompt_text) == 0: if prompt_text is None or len(prompt_text) == 0:
ref_free = True ref_free = True
t0 = ttime() t0 = ttime()
@ -421,7 +421,7 @@ def get_tts_wav(ref_wav_path, prompt_text, prompt_language, text, text_language,
# audio = vq_model.decode(pred_semantic, all_phoneme_ids, refer).detach().cpu().numpy()[0, 0] # audio = vq_model.decode(pred_semantic, all_phoneme_ids, refer).detach().cpu().numpy()[0, 0]
audio = ( audio = (
vq_model.decode( vq_model.decode(
pred_semantic, torch.LongTensor(phones2).to(device).unsqueeze(0), refer pred_semantic, torch.LongTensor(phones2).to(device).unsqueeze(0), refer,speed=speed
) )
.detach() .detach()
.cpu() .cpu()
@ -623,15 +623,17 @@ with gr.Blocks(title="GPT-SoVITS WebUI") as app:
) )
with gr.Row(): with gr.Row():
gr.Markdown(value=i18n("gpt采样参数(无参考文本时不要太低)：")) gr.Markdown(value=i18n("gpt采样参数(无参考文本时不要太低)："))
top_k = gr.Slider(minimum=1,maximum=100,step=1,label=i18n("top_k"),value=5,interactive=True) top_k = gr.Slider(minimum=1,maximum=100,step=1,label=i18n("top_k"),value=10,interactive=True)
top_p = gr.Slider(minimum=0,maximum=1,step=0.05,label=i18n("top_p"),value=1,interactive=True) top_p = gr.Slider(minimum=0,maximum=1,step=0.05,label=i18n("top_p"),value=1,interactive=True)
temperature = gr.Slider(minimum=0,maximum=1,step=0.05,label=i18n("temperature"),value=1,interactive=True) temperature = gr.Slider(minimum=0,maximum=1,step=0.05,label=i18n("temperature"),value=1,interactive=True)
with gr.Row():
speed = gr.Slider(minimum=0.5,maximum=2,step=0.05,label=i18n("speed"),value=1,interactive=True)
inference_button = gr.Button(i18n("合成语音"), variant="primary") inference_button = gr.Button(i18n("合成语音"), variant="primary")
output = gr.Audio(label=i18n("输出的语音")) output = gr.Audio(label=i18n("输出的语音"))
inference_button.click( inference_button.click(
get_tts_wav, get_tts_wav,
[inp_ref, prompt_text, prompt_language, text, text_language, how_to_cut, top_k, top_p, temperature, ref_text_free], [inp_ref, prompt_text, prompt_language, text, text_language, how_to_cut, top_k, top_p, temperature, ref_text_free,speed],
[output], [output],
) )