Spaces:
Running
on
T4
Running
on
T4
Commit
•
406e977
1
Parent(s):
7a62809
Remove tempfile dependency
Browse files- app.py +3 -3
- melo/api.py +2 -2
app.py
CHANGED
@@ -8,9 +8,9 @@ device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
|
8 |
model = TTS(language='EN', device=device)
|
9 |
speaker_ids = model.hps.data.spk2id
|
10 |
def synthesize(speaker, text, speed=1.0, progress=gr.Progress()):
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
with gr.Blocks() as demo:
|
15 |
gr.Markdown('# MeloTTS\n\nAn unofficial demo of [MeloTTS](https://github.com/myshell-ai/MeloTTS) from MyShell AI. MeloTTS is a permissively licensed (MIT) SOTA multi-speaker TTS model.\n\nI am not affiliated with MyShell AI in any way.\n\nThis demo currently only supports English, but the model itself supports other languages.')
|
16 |
with gr.Group():
|
|
|
8 |
model = TTS(language='EN', device=device)
|
9 |
speaker_ids = model.hps.data.spk2id
|
10 |
def synthesize(speaker, text, speed=1.0, progress=gr.Progress()):
|
11 |
+
bio = io.BytesIO()
|
12 |
+
model.tts_to_file(text, speaker_ids[speaker], bio, speed=speed, pbar=progress.tqdm, format='wav')
|
13 |
+
return bio.getvalue()
|
14 |
with gr.Blocks() as demo:
|
15 |
gr.Markdown('# MeloTTS\n\nAn unofficial demo of [MeloTTS](https://github.com/myshell-ai/MeloTTS) from MyShell AI. MeloTTS is a permissively licensed (MIT) SOTA multi-speaker TTS model.\n\nI am not affiliated with MyShell AI in any way.\n\nThis demo currently only supports English, but the model itself supports other languages.')
|
16 |
with gr.Group():
|
melo/api.py
CHANGED
@@ -70,7 +70,7 @@ class TTS(nn.Module):
|
|
70 |
# print(" > ===========================")
|
71 |
return texts
|
72 |
|
73 |
-
def tts_to_file(self, text, speaker_id, output_path=None, sdp_ratio=0.2, noise_scale=0.6, noise_scale_w=0.8, speed=1.0, pbar=None):
|
74 |
language = self.language
|
75 |
texts = self.split_sentences_into_pieces(text, language)
|
76 |
audio_list = []
|
@@ -113,4 +113,4 @@ class TTS(nn.Module):
|
|
113 |
if output_path is None:
|
114 |
return audio
|
115 |
else:
|
116 |
-
soundfile.write(output_path, audio, self.hps.data.sampling_rate)
|
|
|
70 |
# print(" > ===========================")
|
71 |
return texts
|
72 |
|
73 |
+
def tts_to_file(self, text, speaker_id, output_path=None, sdp_ratio=0.2, noise_scale=0.6, noise_scale_w=0.8, speed=1.0, pbar=None, format=None):
|
74 |
language = self.language
|
75 |
texts = self.split_sentences_into_pieces(text, language)
|
76 |
audio_list = []
|
|
|
113 |
if output_path is None:
|
114 |
return audio
|
115 |
else:
|
116 |
+
soundfile.write(output_path, audio, self.hps.data.sampling_rate, format)
|