blackteay/GPT-SoVITS
1
0
Fork 0
mirror of https://github.com/RVC-Boss/GPT-SoVITS.git synced 2025-09-09 22:17:10 +08:00
Code Issues Packages Projects Releases Wiki Activity

Tidy things up

Browse Source
This commit is contained in:
Jarod Mica 2024-12-23 00:24:23 -08:00
parent 251160362f
commit 8292d7bfc2
1 changed files with 9 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
GPT_SoVITS/TTS_infer_pack/TTS.py
View File

@ -1079,6 +1079,9 @@ class TTS:
actual_seed = set_seed(seed) actual_seed = set_seed(seed)
parallel_infer = inputs.get("parallel_infer", True) parallel_infer = inputs.get("parallel_infer", True)
repetition_penalty = inputs.get("repetition_penalty", 1.35) repetition_penalty = inputs.get("repetition_penalty", 1.35)
search_length = inputs.get("search_length", 32000 * 5)
num_zeroes = inputs.get("num_zeroes", 5)
cumulation_amount = inputs.get("cumulation_amount", 50)
if parallel_infer: if parallel_infer:
print(i18n("并行推理模式已开启")) print(i18n("并行推理模式已开启"))
@ -1250,9 +1253,6 @@ class TTS:
crossing_direction = 0 crossing_direction = 0
first_chunk = True first_chunk = True
last_chunk = False last_chunk = False
search_length = 32000*5
num_zeroes = 5
cumulation_amount=50
# Use infer_panel_generator to generate tokens in batches # Use infer_panel_generator to generate tokens in batches
for generated_tokens in self.t2s_model.model.infer_panel_generator( for generated_tokens in self.t2s_model.model.infer_panel_generator(
@ -1313,8 +1313,8 @@ class TTS:
elif center_index >= len(audio_output): elif center_index >= len(audio_output):
raise "Something wrong is going on here, center index issue, greater than audio_output" raise "Something wrong is going on here, center index issue, greater than audio_output"
# How this works: Sequentially progress through audio file in "chunks", starting from the previous zc (zero_crossing) index up and through to the last chunk.
if first_chunk: if first_chunk:
zc_index1, crossing_direction = find_zero_zone( zc_index1, crossing_direction = find_zero_zone(
chunk=audio_output, chunk=audio_output,
start_index=start_index, start_index=start_index,
@ -1432,8 +1432,6 @@ class TTS:
return sr, audio return sr, audio
def speed_change(input_audio:np.ndarray, speed:float, sr:int): def speed_change(input_audio:np.ndarray, speed:float, sr:int):
# 将 NumPy 数组转换为原始 PCM 流 # 将 NumPy 数组转换为原始 PCM 流
raw_audio = input_audio.astype(np.int16).tobytes() raw_audio = input_audio.astype(np.int16).tobytes()