Update lib/tts_infer.py
Browse files- lib/tts_infer.py +15 -2
lib/tts_infer.py
CHANGED
@@ -4,6 +4,7 @@ import gc
|
|
4 |
import torch
|
5 |
from multiprocessing import cpu_count
|
6 |
from lib.modules import VC
|
|
|
7 |
from lib.split_audio import split_silence_nonsilent, adjust_audio_lengths, combine_silence_nonsilent
|
8 |
import edge_tts
|
9 |
import tempfile
|
@@ -109,7 +110,8 @@ def get_model(voice_model):
|
|
109 |
|
110 |
def infer_audio(
|
111 |
model_name,
|
112 |
-
|
|
|
113 |
f0_change=0,
|
114 |
f0_method="rmvpe",
|
115 |
min_pitch="50",
|
@@ -140,6 +142,17 @@ def infer_audio(
|
|
140 |
vc = VC(configs)
|
141 |
pth_path, index_path = get_model(model_name)
|
142 |
vc_data = vc.get_vc(pth_path, protect, 0.5)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
143 |
|
144 |
if split_infer:
|
145 |
inferred_files = []
|
@@ -195,7 +208,7 @@ def infer_audio(
|
|
195 |
else:
|
196 |
inference_info, audio_data, output_path = vc.vc_single(
|
197 |
0,
|
198 |
-
audio_path,
|
199 |
f0_change,
|
200 |
f0_method,
|
201 |
index_path,
|
|
|
4 |
import torch
|
5 |
from multiprocessing import cpu_count
|
6 |
from lib.modules import VC
|
7 |
+
from lib.language_tts import language_dict
|
8 |
from lib.split_audio import split_silence_nonsilent, adjust_audio_lengths, combine_silence_nonsilent
|
9 |
import edge_tts
|
10 |
import tempfile
|
|
|
110 |
|
111 |
def infer_audio(
|
112 |
model_name,
|
113 |
+
text,
|
114 |
+
language_code,
|
115 |
f0_change=0,
|
116 |
f0_method="rmvpe",
|
117 |
min_pitch="50",
|
|
|
142 |
vc = VC(configs)
|
143 |
pth_path, index_path = get_model(model_name)
|
144 |
vc_data = vc.get_vc(pth_path, protect, 0.5)
|
145 |
+
|
146 |
+
|
147 |
+
voice = language_dict.get(language_code, "default_voice")
|
148 |
+
communicate = edge_tts.Communicate(text, voice)
|
149 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
|
150 |
+
tmp_path = tmp_file.name
|
151 |
+
await communicate.save(tmp_path)
|
152 |
+
|
153 |
+
|
154 |
+
|
155 |
+
|
156 |
|
157 |
if split_infer:
|
158 |
inferred_files = []
|
|
|
208 |
else:
|
209 |
inference_info, audio_data, output_path = vc.vc_single(
|
210 |
0,
|
211 |
+
audio_path=tmp_path,
|
212 |
f0_change,
|
213 |
f0_method,
|
214 |
index_path,
|