deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-05 18:09:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix vocoder speech overlap

Browse Source
This commit is contained in:
lyuxiang.lx
2024-08-29 19:10:08 +08:00
parent f1e374a9bb
commit 1d881df8b2
3 changed files with 93 additions and 74 deletions
Download Patch File Download Diff File Expand all files Collapse all files

145
cosyvoice/cli/model.py
View File

@@ -31,18 +31,25 @@ class CosyVoiceModel:
self.flow = flow self.flow = flow
self.hift = hift self.hift = hift
self.token_min_hop_len = 100 self.token_min_hop_len = 100
self.token_max_hop_len = 400 self.token_max_hop_len = 200
self.token_overlap_len = 20 self.token_overlap_len = 20
self.speech_overlap_len = 34 * 256 # mel fade in out
self.window = np.hamming(2 * self.speech_overlap_len) self.mel_overlap_len = 34
self.mel_window = np.hamming(2 * self.mel_overlap_len)
# hift cache
self.mel_cache_len = 20
self.source_cache_len = int(self.mel_cache_len * 256)
# rtf and decoding related
self.stream_scale_factor = 1 self.stream_scale_factor = 1
assert self.stream_scale_factor >= 1, 'stream_scale_factor should be greater than 1, change it according to your actual rtf' assert self.stream_scale_factor >= 1, 'stream_scale_factor should be greater than 1, change it according to your actual rtf'
self.llm_context = torch.cuda.stream(torch.cuda.Stream(self.device)) if torch.cuda.is_available() else nullcontext() self.llm_context = torch.cuda.stream(torch.cuda.Stream(self.device)) if torch.cuda.is_available() else nullcontext()
self.flow_hift_context = torch.cuda.stream(torch.cuda.Stream(self.device)) if torch.cuda.is_available() else nullcontext() self.flow_hift_context = torch.cuda.stream(torch.cuda.Stream(self.device)) if torch.cuda.is_available() else nullcontext()
self.lock = threading.Lock() self.lock = threading.Lock()
# dict used to store session related variable # dict used to store session related variable
self.tts_speech_token = {} self.tts_speech_token_dict = {}
self.llm_end = {} self.llm_end_dict = {}
self.mel_overlap_dict = {}
self.hift_cache_dict = {}
def load(self, llm_model, flow_model, hift_model): def load(self, llm_model, flow_model, hift_model):
self.llm.load_state_dict(torch.load(llm_model, map_location=self.device)) self.llm.load_state_dict(torch.load(llm_model, map_location=self.device))
@@ -64,102 +71,108 @@ class CosyVoiceModel:
self.flow.decoder.estimator = xxx self.flow.decoder.estimator = xxx
self.flow.decoder.session = xxx self.flow.decoder.session = xxx
def llm_job(self, text, text_len, prompt_text, prompt_text_len, llm_prompt_speech_token, llm_prompt_speech_token_len, llm_embedding, this_uuid): def llm_job(self, text, prompt_text, llm_prompt_speech_token, llm_embedding, uuid):
with self.llm_context: with self.llm_context:
for i in self.llm.inference(text=text.to(self.device), for i in self.llm.inference(text=text.to(self.device),
text_len=text_len.to(self.device), text_len=torch.tensor([text.shape[1]], dtype=torch.int32).to(self.device),
prompt_text=prompt_text.to(self.device), prompt_text=prompt_text.to(self.device),
prompt_text_len=prompt_text_len.to(self.device), prompt_text_len=torch.tensor([prompt_text.shape[1]], dtype=torch.int32).to(self.device),
prompt_speech_token=llm_prompt_speech_token.to(self.device), prompt_speech_token=llm_prompt_speech_token.to(self.device),
prompt_speech_token_len=llm_prompt_speech_token_len.to(self.device), prompt_speech_token_len=torch.tensor([llm_prompt_speech_token.shape[1]], dtype=torch.int32).to(self.device),
embedding=llm_embedding.to(self.device).half(), embedding=llm_embedding.to(self.device).half(),
sampling=25, sampling=25,
max_token_text_ratio=30, max_token_text_ratio=30,
min_token_text_ratio=3): min_token_text_ratio=3):
self.tts_speech_token[this_uuid].append(i) self.tts_speech_token_dict[uuid].append(i)
self.llm_end[this_uuid] = True self.llm_end_dict[uuid] = True
def token2wav(self, token, prompt_token, prompt_token_len, prompt_feat, prompt_feat_len, embedding): def token2wav(self, token, prompt_token, prompt_feat, embedding, uuid, finalize=False):
with self.flow_hift_context: with self.flow_hift_context:
tts_mel = self.flow.inference(token=token.to(self.device), tts_mel = self.flow.inference(token=token.to(self.device),
token_len=torch.tensor([token.size(1)], dtype=torch.int32).to(self.device), token_len=torch.tensor([token.shape[1]], dtype=torch.int32).to(self.device),
prompt_token=prompt_token.to(self.device), prompt_token=prompt_token.to(self.device),
prompt_token_len=prompt_token_len.to(self.device), prompt_token_len=torch.tensor([prompt_token.shape[1]], dtype=torch.int32).to(self.device),
prompt_feat=prompt_feat.to(self.device), prompt_feat=prompt_feat.to(self.device),
prompt_feat_len=prompt_feat_len.to(self.device), prompt_feat_len=torch.tensor([prompt_feat.shape[1]], dtype=torch.int32).to(self.device),
embedding=embedding.to(self.device)) embedding=embedding.to(self.device))
tts_speech = self.hift.inference(mel=tts_mel).cpu() # mel overlap fade in out
if self.mel_overlap_dict[uuid] is not None:
tts_mel = fade_in_out(tts_mel, self.mel_overlap_dict[uuid], self.mel_window)
# append hift cache
if self.hift_cache_dict[uuid] is not None:
hift_cache_mel, hift_cache_source = self.hift_cache_dict[uuid]['mel'], self.hift_cache_dict[uuid]['source']
tts_mel = torch.concat([hift_cache_mel, tts_mel], dim=2)
else:
hift_cache_source = torch.zeros(1, 1, 0)
# keep overlap mel and hift cache
if finalize is False:
self.mel_overlap_dict[uuid] = tts_mel[:, :, -self.mel_overlap_len:]
tts_mel = tts_mel[:, :, :-self.mel_overlap_len]
tts_speech, tts_source = self.hift.inference(mel=tts_mel, cache_source=hift_cache_source)
self.hift_cache_dict[uuid] = {'source': tts_source[:, :, -self.source_cache_len:], 'mel': tts_mel[:, :, -self.mel_cache_len:]}
tts_speech = tts_speech[:, :-self.source_cache_len]
else:
tts_speech, tts_source = self.hift.inference(mel=tts_mel, cache_source=hift_cache_source)
return tts_speech return tts_speech
def inference(self, text, text_len, flow_embedding, llm_embedding=torch.zeros(0, 192), def inference(self, text, flow_embedding, llm_embedding=torch.zeros(0, 192),
prompt_text=torch.zeros(1, 0, dtype=torch.int32), prompt_text_len=torch.zeros(1, dtype=torch.int32), prompt_text=torch.zeros(1, 0, dtype=torch.int32),
llm_prompt_speech_token=torch.zeros(1, 0, dtype=torch.int32), llm_prompt_speech_token_len=torch.zeros(1, dtype=torch.int32), llm_prompt_speech_token=torch.zeros(1, 0, dtype=torch.int32),
flow_prompt_speech_token=torch.zeros(1, 0, dtype=torch.int32), flow_prompt_speech_token_len=torch.zeros(1, dtype=torch.int32), flow_prompt_speech_token=torch.zeros(1, 0, dtype=torch.int32),
prompt_speech_feat=torch.zeros(1, 0, 80), prompt_speech_feat_len=torch.zeros(1, dtype=torch.int32), stream=False): prompt_speech_feat=torch.zeros(1, 0, 80), stream=False, **kwargs):
# this_uuid is used to track variables related to this inference thread # this_uuid is used to track variables related to this inference thread
this_uuid = str(uuid.uuid1()) this_uuid = str(uuid.uuid1())
with self.lock: with self.lock:
self.tts_speech_token[this_uuid], self.llm_end[this_uuid] = [], False self.tts_speech_token_dict[this_uuid], self.llm_end_dict[this_uuid], self.mel_overlap_dict[this_uuid], self.hift_cache_dict[this_uuid] = [], False, None, None
p = threading.Thread(target=self.llm_job, args=(text.to(self.device), text_len.to(self.device), prompt_text.to(self.device), prompt_text_len.to(self.device), p = threading.Thread(target=self.llm_job, args=(text, prompt_text, llm_prompt_speech_token, llm_embedding, this_uuid))
llm_prompt_speech_token.to(self.device), llm_prompt_speech_token_len.to(self.device), llm_embedding.to(self.device), this_uuid))
p.start() p.start()
p.join()
if stream is True: if stream is True:
cache_speech, cache_token, token_hop_len = None, None, self.token_min_hop_len token_hop_len = self.token_min_hop_len
while True: while True:
time.sleep(0.1) time.sleep(0.1)
if len(self.tts_speech_token[this_uuid]) >= token_hop_len + self.token_overlap_len: if len(self.tts_speech_token_dict[this_uuid]) >= token_hop_len + self.token_overlap_len:
this_tts_speech_token = torch.concat(self.tts_speech_token[this_uuid][:token_hop_len + self.token_overlap_len], dim=1) this_tts_speech_token = torch.concat(self.tts_speech_token_dict[this_uuid][:token_hop_len + self.token_overlap_len], dim=1)
with self.flow_hift_context: with self.flow_hift_context:
this_tts_speech = self.token2wav(token=this_tts_speech_token, this_tts_speech = self.token2wav(token=this_tts_speech_token,
prompt_token=flow_prompt_speech_token.to(self.device), prompt_token=flow_prompt_speech_token,
prompt_token_len=flow_prompt_speech_token_len.to(self.device), prompt_feat=prompt_speech_feat,
prompt_feat=prompt_speech_feat.to(self.device), embedding=flow_embedding,
prompt_feat_len=prompt_speech_feat_len.to(self.device), uuid=this_uuid,
embedding=flow_embedding.to(self.device)) finalize=False)
# fade in/out if necessary yield {'tts_speech': this_tts_speech.cpu()}
if cache_speech is not None:
this_tts_speech = fade_in_out(this_tts_speech, cache_speech, self.window)
yield {'tts_speech': this_tts_speech[:, :-self.speech_overlap_len]}
cache_speech = this_tts_speech[:, -self.speech_overlap_len:]
cache_token = self.tts_speech_token[this_uuid][:token_hop_len]
with self.lock: with self.lock:
self.tts_speech_token[this_uuid] = self.tts_speech_token[this_uuid][token_hop_len:] self.tts_speech_token_dict[this_uuid] = self.tts_speech_token_dict[this_uuid][token_hop_len:]
# increase token_hop_len for better speech quality # increase token_hop_len for better speech quality
token_hop_len = min(self.token_max_hop_len, int(token_hop_len * self.stream_scale_factor)) token_hop_len = min(self.token_max_hop_len, int(token_hop_len * self.stream_scale_factor))
if self.llm_end[this_uuid] is True and len(self.tts_speech_token[this_uuid]) < token_hop_len + self.token_overlap_len: if self.llm_end_dict[this_uuid] is True and len(self.tts_speech_token_dict[this_uuid]) < token_hop_len + self.token_overlap_len:
break break
p.join() # p.join()
# deal with remain tokens, make sure inference remain token len equals token_hop_len when cache_speech is not None # deal with remain tokens, make sure inference remain token len equals token_hop_len when cache_speech is not None
this_tts_speech_token = torch.concat(self.tts_speech_token[this_uuid], dim=1) this_tts_speech_token = torch.concat(self.tts_speech_token_dict[this_uuid], dim=1)
if this_tts_speech_token.shape[1] < self.token_min_hop_len + self.token_overlap_len and cache_token is not None:
cache_token_len = self.token_min_hop_len + self.token_overlap_len - this_tts_speech_token.shape[1]
this_tts_speech_token = torch.concat([torch.concat(cache_token[-cache_token_len:], dim=1), this_tts_speech_token], dim=1)
else:
cache_token_len = 0
with self.flow_hift_context: with self.flow_hift_context:
this_tts_speech = self.token2wav(token=this_tts_speech_token, this_tts_speech = self.token2wav(token=this_tts_speech_token,
prompt_token=flow_prompt_speech_token.to(self.device), prompt_token=flow_prompt_speech_token,
prompt_token_len=flow_prompt_speech_token_len.to(self.device), prompt_feat=prompt_speech_feat,
prompt_feat=prompt_speech_feat.to(self.device), embedding=flow_embedding,
prompt_feat_len=prompt_speech_feat_len.to(self.device), uuid=this_uuid,
embedding=flow_embedding.to(self.device)) finalize=True)
this_tts_speech = this_tts_speech[:, int(cache_token_len / this_tts_speech_token.shape[1] * this_tts_speech.shape[1]):] yield {'tts_speech': this_tts_speech.cpu()}
if cache_speech is not None:
this_tts_speech = fade_in_out(this_tts_speech, cache_speech, self.window)
yield {'tts_speech': this_tts_speech}
else: else:
# deal with all tokens # deal with all tokens
p.join() # p.join()
this_tts_speech_token = torch.concat(self.tts_speech_token[this_uuid], dim=1) this_tts_speech_token = torch.concat(self.tts_speech_token_dict[this_uuid], dim=1)
with self.flow_hift_context: with self.flow_hift_context:
this_tts_speech = self.token2wav(token=this_tts_speech_token, this_tts_speech = self.token2wav(token=this_tts_speech_token,
prompt_token=flow_prompt_speech_token.to(self.device), prompt_token=flow_prompt_speech_token,
prompt_token_len=flow_prompt_speech_token_len.to(self.device), prompt_feat=prompt_speech_feat,
prompt_feat=prompt_speech_feat.to(self.device), embedding=flow_embedding,
prompt_feat_len=prompt_speech_feat_len.to(self.device), uuid=this_uuid,
embedding=flow_embedding.to(self.device)) finalize=True)
yield {'tts_speech': this_tts_speech} yield {'tts_speech': this_tts_speech.cpu()}
with self.lock: with self.lock:
self.tts_speech_token.pop(this_uuid) self.tts_speech_token_dict.pop(this_uuid)
self.llm_end.pop(this_uuid) self.llm_end_dict.pop(this_uuid)
self.mel_overlap_dict.pop(this_uuid)
self.hift_cache_dict.pop(this_uuid)
torch.cuda.synchronize() torch.cuda.synchronize()

12
cosyvoice/hifigan/generator.py
View File

@@ -335,10 +335,14 @@ class HiFTGenerator(nn.Module):
inverse_transform = torch.istft(torch.complex(real, img), self.istft_params["n_fft"], self.istft_params["hop_len"], self.istft_params["n_fft"], window=self.stft_window.to(magnitude.device)) inverse_transform = torch.istft(torch.complex(real, img), self.istft_params["n_fft"], self.istft_params["hop_len"], self.istft_params["n_fft"], window=self.stft_window.to(magnitude.device))
return inverse_transform return inverse_transform
def forward(self, x: torch.Tensor) -> torch.Tensor: def forward(self, x: torch.Tensor, cache_source: torch.Tensor = torch.zeros(1, 1, 0)) -> torch.Tensor:
f0 = self.f0_predictor(x) f0 = self.f0_predictor(x)
s = self._f02source(f0) s = self._f02source(f0)
# use cache_source to avoid glitch
if cache_source.shape[2] == 0:
s[:, :, :cache_source.shape[2]] = cache_source
s_stft_real, s_stft_imag = self._stft(s.squeeze(1)) s_stft_real, s_stft_imag = self._stft(s.squeeze(1))
s_stft = torch.cat([s_stft_real, s_stft_imag], dim=1) s_stft = torch.cat([s_stft_real, s_stft_imag], dim=1)
@@ -370,7 +374,7 @@ class HiFTGenerator(nn.Module):
x = self._istft(magnitude, phase) x = self._istft(magnitude, phase)
x = torch.clamp(x, -self.audio_limit, self.audio_limit) x = torch.clamp(x, -self.audio_limit, self.audio_limit)
return x return x, s
def remove_weight_norm(self): def remove_weight_norm(self):
print('Removing weight norm...') print('Removing weight norm...')
@@ -387,5 +391,5 @@ class HiFTGenerator(nn.Module):
l.remove_weight_norm() l.remove_weight_norm()
@torch.inference_mode() @torch.inference_mode()
def inference(self, mel: torch.Tensor) -> torch.Tensor: def inference(self, mel: torch.Tensor, cache_source: torch.Tensor = torch.zeros(1, 1, 0)) -> torch.Tensor:
return self.forward(x=mel) return self.forward(x=mel, cache_source=cache_source)

10
cosyvoice/utils/common.py
View File

@@ -131,7 +131,9 @@ def random_sampling(weighted_scores, decoded_tokens, sampling):
top_ids = weighted_scores.softmax(dim=0).multinomial(1, replacement=True) top_ids = weighted_scores.softmax(dim=0).multinomial(1, replacement=True)
return top_ids return top_ids
def fade_in_out(fade_in_speech, fade_out_speech, window): def fade_in_out(fade_in_mel, fade_out_mel, window):
speech_overlap_len = int(window.shape[0] / 2) device = fade_in_mel.device
fade_in_speech[:, :speech_overlap_len] = fade_in_speech[:, :speech_overlap_len] * window[:speech_overlap_len] + fade_out_speech[:, -speech_overlap_len:] * window[speech_overlap_len:] fade_in_mel, fade_out_mel = fade_in_mel.cpu(), fade_out_mel.cpu()
return fade_in_speech mel_overlap_len = int(window.shape[0] / 2)
fade_in_mel[:, :, :mel_overlap_len] = fade_in_mel[:, :, :mel_overlap_len] * window[:mel_overlap_len] + fade_out_mel[:, :, -mel_overlap_len:] * window[mel_overlap_len:]
return fade_in_mel.to(device)