deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-05 09:59:23 +08:00
Code Issues Packages Projects Releases Wiki Activity

update export_codec_vllm

Browse Source
This commit is contained in:
雾聪
2025-02-26 20:25:14 +08:00
parent f280558bcb
commit 54e9384fb1
2 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
cosyvoice/cli/model.py
View File

@@ -331,6 +331,7 @@ class CosyVoice2Model(CosyVoiceModel):
pad_vocab_size = ((vocab_size + pad_to - 1) // pad_to) * pad_to pad_vocab_size = ((vocab_size + pad_to - 1) // pad_to) * pad_to
dtype = torch.bfloat16 dtype = torch.bfloat16
# lm_head
new_lm_head = nn.Linear(in_features=feature_size, out_features=pad_vocab_size, bias=True) new_lm_head = nn.Linear(in_features=feature_size, out_features=pad_vocab_size, bias=True)
with torch.no_grad(): with torch.no_grad():
new_lm_head.weight[:vocab_size] = self.llm.llm_decoder.weight new_lm_head.weight[:vocab_size] = self.llm.llm_decoder.weight
@@ -339,6 +340,8 @@ class CosyVoice2Model(CosyVoiceModel):
new_lm_head.bias[vocab_size:] = 0 new_lm_head.bias[vocab_size:] = 0
self.llm.llm.model.lm_head = new_lm_head self.llm.llm.model.lm_head = new_lm_head
new_codec_embed = nn.Linear(in_features=feature_size, out_features=pad_vocab_size) new_codec_embed = nn.Linear(in_features=feature_size, out_features=pad_vocab_size)
# embed_tokens
embed_tokens = self.llm.llm.model.model.embed_tokens
with torch.no_grad(): with torch.no_grad():
new_codec_embed.weight[:vocab_size] = self.llm.speech_embedding.weight new_codec_embed.weight[:vocab_size] = self.llm.speech_embedding.weight
new_codec_embed.weight[vocab_size:] = 0 new_codec_embed.weight[vocab_size:] = 0
@@ -356,6 +359,7 @@ class CosyVoice2Model(CosyVoiceModel):
self.llm.llm.model.save_pretrained(model_path) self.llm.llm.model.save_pretrained(model_path)
self.llm.llm.model.config.vocab_size = tmp_vocab_size self.llm.llm.model.config.vocab_size = tmp_vocab_size
self.llm.llm.model.config.tie_word_embeddings = tmp_tie_embedding self.llm.llm.model.config.tie_word_embeddings = tmp_tie_embedding
self.llm.llm.model.set_input_embeddings(embed_tokens)
def token2wav(self, token, prompt_token, prompt_feat, embedding, uuid, token_offset, finalize=False, speed=1.0): def token2wav(self, token, prompt_token, prompt_feat, embedding, uuid, token_offset, finalize=False, speed=1.0):
tts_mel, _ = self.flow.inference(token=token.to(self.device), tts_mel, _ = self.flow.inference(token=token.to(self.device),

4
cosyvoice/llm/llm.py
View File

@@ -353,12 +353,14 @@ class Qwen2LM(TransformerLM):
if str(request_output.request_id) != str(request_id): if str(request_output.request_id) != str(request_id):
continue continue
if not request_output.finished: if not request_output.finished:
print(f"Partial request output: {request_output}") # print(f"Partial request output: {request_output}")
out_token = list(request_output.outputs[0].token_ids)[-1] out_token = list(request_output.outputs[0].token_ids)[-1]
yield out_token yield out_token
out_token_ids.append(out_token) out_token_ids.append(out_token)
else: else:
break break
if not vllm_codec_engine.has_unfinished_requests():
break
@torch.inference_mode() @torch.inference_mode()
def inference_bistream( def inference_bistream(