mirror of
https://github.com/RVC-Boss/GPT-SoVITS.git
synced 2025-04-29 22:10:21 +08:00
修复了OutOfMemoryError时，显存无法释放的问题
This commit is contained in:
chasonjiang
parent
f2cbc826c7
commit
d60d8ea3fb
@ -2,6 +2,7 @@ from copy import deepcopy
|
|||||||
import math
|
import math
|
||||||
import os, sys
|
import os, sys
|
||||||
import random
|
import random
|
||||||
|
import traceback
|
||||||
now_dir = os.getcwd()
|
now_dir = os.getcwd()
|
||||||
sys.path.append(now_dir)
|
sys.path.append(now_dir)
|
||||||
import ffmpeg
|
import ffmpeg
|
||||||
@ -48,7 +49,17 @@ custom:
|
|||||||
|
|
||||||
"""
|
"""
|
||||||
|
|
||||||
|
# def set_seed(seed):
|
||||||
|
# random.seed(seed)
|
||||||
|
# os.environ['PYTHONHASHSEED'] = str(seed)
|
||||||
|
# np.random.seed(seed)
|
||||||
|
# torch.manual_seed(seed)
|
||||||
|
# torch.cuda.manual_seed(seed)
|
||||||
|
# torch.cuda.manual_seed_all(seed)
|
||||||
|
# torch.backends.cudnn.deterministic = True
|
||||||
|
# torch.backends.cudnn.benchmark = False
|
||||||
|
# torch.backends.cudnn.enabled = True
|
||||||
|
# set_seed(1234)
|
||||||
|
|
||||||
class TTS_Config:
|
class TTS_Config:
|
||||||
default_configs={
|
default_configs={
|
||||||
@ -630,125 +641,141 @@ class TTS:
|
|||||||
split_bucket=split_bucket
|
split_bucket=split_bucket
|
||||||
)
|
)
|
||||||
t2 = ttime()
|
t2 = ttime()
|
||||||
|
try:
|
||||||
|
print("############ 推理 ############")
|
||||||
|
###### inference ######
|
||||||
|
t_34 = 0.0
|
||||||
|
t_45 = 0.0
|
||||||
|
audio = []
|
||||||
|
for item in data:
|
||||||
|
t3 = ttime()
|
||||||
|
batch_phones = item["phones"]
|
||||||
|
batch_phones_len = item["phones_len"]
|
||||||
|
all_phoneme_ids = item["all_phones"]
|
||||||
|
all_phoneme_lens = item["all_phones_len"]
|
||||||
|
all_bert_features = item["all_bert_features"]
|
||||||
|
norm_text = item["norm_text"]
|
||||||
|
|
||||||
print("############ 推理 ############")
|
# batch_phones = batch_phones.to(self.configs.device)
|
||||||
###### inference ######
|
batch_phones_len = batch_phones_len.to(self.configs.device)
|
||||||
t_34 = 0.0
|
all_phoneme_ids = all_phoneme_ids.to(self.configs.device)
|
||||||
t_45 = 0.0
|
all_phoneme_lens = all_phoneme_lens.to(self.configs.device)
|
||||||
audio = []
|
all_bert_features = all_bert_features.to(self.configs.device)
|
||||||
for item in data:
|
if self.configs.is_half:
|
||||||
t3 = ttime()
|
all_bert_features = all_bert_features.half()
|
||||||
batch_phones = item["phones"]
|
|
||||||
batch_phones_len = item["phones_len"]
|
|
||||||
all_phoneme_ids = item["all_phones"]
|
|
||||||
all_phoneme_lens = item["all_phones_len"]
|
|
||||||
all_bert_features = item["all_bert_features"]
|
|
||||||
norm_text = item["norm_text"]
|
|
||||||
|
|
||||||
# batch_phones = batch_phones.to(self.configs.device)
|
print(i18n("前端处理后的文本(每句):"), norm_text)
|
||||||
batch_phones_len = batch_phones_len.to(self.configs.device)
|
if no_prompt_text :
|
||||||
all_phoneme_ids = all_phoneme_ids.to(self.configs.device)
|
prompt = None
|
||||||
all_phoneme_lens = all_phoneme_lens.to(self.configs.device)
|
else:
|
||||||
all_bert_features = all_bert_features.to(self.configs.device)
|
prompt = self.prompt_cache["prompt_semantic"].expand(all_phoneme_ids.shape[0], -1).to(self.configs.device)
|
||||||
if self.configs.is_half:
|
|
||||||
all_bert_features = all_bert_features.half()
|
|
||||||
|
|
||||||
print(i18n("前端处理后的文本(每句):"), norm_text)
|
with torch.no_grad():
|
||||||
if no_prompt_text :
|
pred_semantic_list, idx_list = self.t2s_model.model.infer_panel(
|
||||||
prompt = None
|
all_phoneme_ids,
|
||||||
else:
|
all_phoneme_lens,
|
||||||
prompt = self.prompt_cache["prompt_semantic"].expand(all_phoneme_ids.shape[0], -1).to(self.configs.device)
|
prompt,
|
||||||
|
all_bert_features,
|
||||||
|
# prompt_phone_len=ph_offset,
|
||||||
|
top_k=top_k,
|
||||||
|
top_p=top_p,
|
||||||
|
temperature=temperature,
|
||||||
|
early_stop_num=self.configs.hz * self.configs.max_sec,
|
||||||
|
)
|
||||||
|
t4 = ttime()
|
||||||
|
t_34 += t4 - t3
|
||||||
|
|
||||||
with torch.no_grad():
|
refer_audio_spepc:torch.Tensor = self.prompt_cache["refer_spepc"]\
|
||||||
pred_semantic_list, idx_list = self.t2s_model.model.infer_panel(
|
.to(dtype=self.precison, device=self.configs.device)
|
||||||
all_phoneme_ids,
|
|
||||||
all_phoneme_lens,
|
|
||||||
prompt,
|
|
||||||
all_bert_features,
|
|
||||||
# prompt_phone_len=ph_offset,
|
|
||||||
top_k=top_k,
|
|
||||||
top_p=top_p,
|
|
||||||
temperature=temperature,
|
|
||||||
early_stop_num=self.configs.hz * self.configs.max_sec,
|
|
||||||
)
|
|
||||||
t4 = ttime()
|
|
||||||
t_34 += t4 - t3
|
|
||||||
|
|
||||||
refer_audio_spepc:torch.Tensor = self.prompt_cache["refer_spepc"]\
|
batch_audio_fragment = []
|
||||||
.to(dtype=self.precison, device=self.configs.device)
|
|
||||||
|
|
||||||
batch_audio_fragment = []
|
# ## vits并行推理 method 1
|
||||||
|
# pred_semantic_list = [item[-idx:] for item, idx in zip(pred_semantic_list, idx_list)]
|
||||||
|
# pred_semantic_len = torch.LongTensor([item.shape[0] for item in pred_semantic_list]).to(self.configs.device)
|
||||||
|
# pred_semantic = self.batch_sequences(pred_semantic_list, axis=0, pad_value=0).unsqueeze(0)
|
||||||
|
# max_len = 0
|
||||||
|
# for i in range(0, len(batch_phones)):
|
||||||
|
# max_len = max(max_len, batch_phones[i].shape[-1])
|
||||||
|
# batch_phones = self.batch_sequences(batch_phones, axis=0, pad_value=0, max_length=max_len)
|
||||||
|
# batch_phones = batch_phones.to(self.configs.device)
|
||||||
|
# batch_audio_fragment = (self.vits_model.batched_decode(
|
||||||
|
# pred_semantic, pred_semantic_len, batch_phones, batch_phones_len,refer_audio_spepc
|
||||||
|
# ))
|
||||||
|
|
||||||
# ## vits并行推理 method 1
|
# ## vits并行推理 method 2
|
||||||
# pred_semantic_list = [item[-idx:] for item, idx in zip(pred_semantic_list, idx_list)]
|
pred_semantic_list = [item[-idx:] for item, idx in zip(pred_semantic_list, idx_list)]
|
||||||
# pred_semantic_len = torch.LongTensor([item.shape[0] for item in pred_semantic_list]).to(self.configs.device)
|
upsample_rate = math.prod(self.vits_model.upsample_rates)
|
||||||
# pred_semantic = self.batch_sequences(pred_semantic_list, axis=0, pad_value=0).unsqueeze(0)
|
audio_frag_idx = [pred_semantic_list[i].shape[0]*2*upsample_rate for i in range(0, len(pred_semantic_list))]
|
||||||
# max_len = 0
|
audio_frag_end_idx = [ sum(audio_frag_idx[:i+1]) for i in range(0, len(audio_frag_idx))]
|
||||||
# for i in range(0, len(batch_phones)):
|
all_pred_semantic = torch.cat(pred_semantic_list).unsqueeze(0).unsqueeze(0).to(self.configs.device)
|
||||||
# max_len = max(max_len, batch_phones[i].shape[-1])
|
_batch_phones = torch.cat(batch_phones).unsqueeze(0).to(self.configs.device)
|
||||||
# batch_phones = self.batch_sequences(batch_phones, axis=0, pad_value=0, max_length=max_len)
|
_batch_audio_fragment = (self.vits_model.decode(
|
||||||
# batch_phones = batch_phones.to(self.configs.device)
|
all_pred_semantic, _batch_phones,refer_audio_spepc
|
||||||
# batch_audio_fragment = (self.vits_model.batched_decode(
|
).detach()[0, 0, :])
|
||||||
# pred_semantic, pred_semantic_len, batch_phones, batch_phones_len,refer_audio_spepc
|
audio_frag_end_idx.insert(0, 0)
|
||||||
# ))
|
batch_audio_fragment= [_batch_audio_fragment[audio_frag_end_idx[i-1]:audio_frag_end_idx[i]] for i in range(1, len(audio_frag_end_idx))]
|
||||||
|
|
||||||
# ## vits并行推理 method 2
|
|
||||||
pred_semantic_list = [item[-idx:] for item, idx in zip(pred_semantic_list, idx_list)]
|
|
||||||
upsample_rate = math.prod(self.vits_model.upsample_rates)
|
|
||||||
audio_frag_idx = [pred_semantic_list[i].shape[0]*2*upsample_rate for i in range(0, len(pred_semantic_list))]
|
|
||||||
audio_frag_end_idx = [ sum(audio_frag_idx[:i+1]) for i in range(0, len(audio_frag_idx))]
|
|
||||||
all_pred_semantic = torch.cat(pred_semantic_list).unsqueeze(0).unsqueeze(0).to(self.configs.device)
|
|
||||||
_batch_phones = torch.cat(batch_phones).unsqueeze(0).to(self.configs.device)
|
|
||||||
_batch_audio_fragment = (self.vits_model.decode(
|
|
||||||
all_pred_semantic, _batch_phones,refer_audio_spepc
|
|
||||||
).detach()[0, 0, :])
|
|
||||||
audio_frag_end_idx.insert(0, 0)
|
|
||||||
batch_audio_fragment= [_batch_audio_fragment[audio_frag_end_idx[i-1]:audio_frag_end_idx[i]] for i in range(1, len(audio_frag_end_idx))]
|
|
||||||
|
|
||||||
|
|
||||||
# ## vits串行推理
|
# ## vits串行推理
|
||||||
# for i, idx in enumerate(idx_list):
|
# for i, idx in enumerate(idx_list):
|
||||||
# phones = batch_phones[i].unsqueeze(0).to(self.configs.device)
|
# phones = batch_phones[i].unsqueeze(0).to(self.configs.device)
|
||||||
# _pred_semantic = (pred_semantic_list[i][-idx:].unsqueeze(0).unsqueeze(0)) # .unsqueeze(0)#mq要多unsqueeze一次
|
# _pred_semantic = (pred_semantic_list[i][-idx:].unsqueeze(0).unsqueeze(0)) # .unsqueeze(0)#mq要多unsqueeze一次
|
||||||
# audio_fragment =(self.vits_model.decode(
|
# audio_fragment =(self.vits_model.decode(
|
||||||
# _pred_semantic, phones, refer_audio_spepc
|
# _pred_semantic, phones, refer_audio_spepc
|
||||||
# ).detach()[0, 0, :])
|
# ).detach()[0, 0, :])
|
||||||
# batch_audio_fragment.append(
|
# batch_audio_fragment.append(
|
||||||
# audio_fragment
|
# audio_fragment
|
||||||
# ) ###试试重建不带上prompt部分
|
# ) ###试试重建不带上prompt部分
|
||||||
|
|
||||||
t5 = ttime()
|
t5 = ttime()
|
||||||
t_45 += t5 - t4
|
t_45 += t5 - t4
|
||||||
if return_fragment:
|
if return_fragment:
|
||||||
print("%.3f\t%.3f\t%.3f\t%.3f" % (t1 - t0, t2 - t1, t4 - t3, t5 - t4))
|
print("%.3f\t%.3f\t%.3f\t%.3f" % (t1 - t0, t2 - t1, t4 - t3, t5 - t4))
|
||||||
yield self.audio_postprocess([batch_audio_fragment],
|
yield self.audio_postprocess([batch_audio_fragment],
|
||||||
|
self.configs.sampling_rate,
|
||||||
|
batch_index_list,
|
||||||
|
speed_factor,
|
||||||
|
split_bucket)
|
||||||
|
else:
|
||||||
|
audio.append(batch_audio_fragment)
|
||||||
|
|
||||||
|
if self.stop_flag:
|
||||||
|
yield self.configs.sampling_rate, np.zeros(int(self.configs.sampling_rate * 0.3),
|
||||||
|
dtype=np.int16)
|
||||||
|
return
|
||||||
|
|
||||||
|
if not return_fragment:
|
||||||
|
print("%.3f\t%.3f\t%.3f\t%.3f" % (t1 - t0, t2 - t1, t_34, t_45))
|
||||||
|
yield self.audio_postprocess(audio,
|
||||||
self.configs.sampling_rate,
|
self.configs.sampling_rate,
|
||||||
batch_index_list,
|
batch_index_list,
|
||||||
speed_factor,
|
speed_factor,
|
||||||
split_bucket)
|
split_bucket)
|
||||||
else:
|
except Exception as e:
|
||||||
audio.append(batch_audio_fragment)
|
traceback.print_exc()
|
||||||
|
# 必须返回一个空音频, 否则会导致显存不释放。
|
||||||
if self.stop_flag:
|
yield self.configs.sampling_rate, np.zeros(int(self.configs.sampling_rate),
|
||||||
yield self.configs.sampling_rate, np.zeros(int(self.configs.sampling_rate * 0.3),
|
dtype=np.int16)
|
||||||
dtype=np.int16)
|
# 重置模型, 否则会导致显存释放不完全。
|
||||||
return
|
del self.t2s_model
|
||||||
|
del self.vits_model
|
||||||
if not return_fragment:
|
self.t2s_model = None
|
||||||
print("%.3f\t%.3f\t%.3f\t%.3f" % (t1 - t0, t2 - t1, t_34, t_45))
|
self.vits_model = None
|
||||||
yield self.audio_postprocess(audio,
|
self.init_t2s_weights(self.configs.t2s_weights_path)
|
||||||
self.configs.sampling_rate,
|
self.init_vits_weights(self.configs.vits_weights_path)
|
||||||
batch_index_list,
|
finally:
|
||||||
speed_factor,
|
self.empty_cache()
|
||||||
split_bucket)
|
|
||||||
|
|
||||||
|
def empty_cache(self):
|
||||||
try:
|
try:
|
||||||
torch.cuda.empty_cache()
|
if str(self.configs.device) == "cuda":
|
||||||
|
torch.cuda.empty_cache()
|
||||||
|
elif str(self.configs.device) == "mps":
|
||||||
|
torch.mps.empty_cache()
|
||||||
except:
|
except:
|
||||||
pass
|
pass
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
def audio_postprocess(self,
|
def audio_postprocess(self,
|
||||||
audio:List[torch.Tensor],
|
audio:List[torch.Tensor],
|
||||||
sr:int,
|
sr:int,
|
||||||
|
|||||||
Loading…
x
Reference in New Issue
Block a user