blackteay/GPT-SoVITS
1
0
Fork 0
mirror of https://github.com/RVC-Boss/GPT-SoVITS.git synced 2025-08-20 10:18:32 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix AttributeError when prompt_cache['refer_spec'][0] is a tuple (#2428)

Browse Source 
Co-authored-by: tzrain <tz_rain@foxmail.com>
This commit is contained in:
Jialiang Zhu 2025-06-05 10:55:21 +08:00 committed by GitHub
parent a080e19f91
commit 035dcbad03
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 8 additions and 2 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

@ -1407,7 +1407,10 @@ class TTS:
):
prompt_semantic_tokens = self.prompt_cache["prompt_semantic"].unsqueeze(0).unsqueeze(0).to(self.configs.device)
prompt_phones = torch.LongTensor(self.prompt_cache["phones"]).unsqueeze(0).to(self.configs.device)
refer_audio_spec = self.prompt_cache["refer_spec"][0].to(dtype=self.precision, device=self.configs.device)
raw_entry = self.prompt_cache["refer_spec"][0]
if isinstance(raw_entry, tuple):
raw_entry = raw_entry[0]
refer_audio_spec = raw_entry.to(dtype=self.precision,device=self.configs.device)
fea_ref, ge = self.vits_model.decode_encp(prompt_semantic_tokens, prompt_phones, refer_audio_spec)
ref_audio: torch.Tensor = self.prompt_cache["raw_audio"]
@ -1474,7 +1477,10 @@ class TTS:
) -> List[torch.Tensor]:
prompt_semantic_tokens = self.prompt_cache["prompt_semantic"].unsqueeze(0).unsqueeze(0).to(self.configs.device)
prompt_phones = torch.LongTensor(self.prompt_cache["phones"]).unsqueeze(0).to(self.configs.device)
refer_audio_spec = self.prompt_cache["refer_spec"][0].to(dtype=self.precision, device=self.configs.device)
raw_entry = self.prompt_cache["refer_spec"][0]
if isinstance(raw_entry, tuple):
raw_entry = raw_entry[0]
refer_audio_spec = raw_entry.to(dtype=self.precision,device=self.configs.device)
fea_ref, ge = self.vits_model.decode_encp(prompt_semantic_tokens, prompt_phones, refer_audio_spec)
ref_audio: torch.Tensor = self.prompt_cache["raw_audio"]