Search results

流式输出的克隆音色效果差别很大，电音很重，请问是什么问题

Status: Open.

#1366 In FunAudioLLM/CosyVoice;

· Chenruxin opened on Jun 11, 2025

CosyVoice-300M-Instruct 模型选择中文男性声音生成的视频仍然是女性声音

Status: Open.

#1365 In FunAudioLLM/CosyVoice;

· F5Psyche opened on Jun 11, 2025

首先cv2效果很好，但是速度有点慢，vllm 不支持windows，windows有什么办法可以提速吗？

Status: Open.

#1364 In FunAudioLLM/CosyVoice;

· theajack opened on Jun 10, 2025

同样的配置，4090推理速度比T4,慢3倍？

Status: Open.

#1362 In FunAudioLLM/CosyVoice;

· soup-zhang opened on Jun 10, 2025

流式flow+TensorRT推理，偶现句子变成静音段。

Status: Open.

#1361 In FunAudioLLM/CosyVoice;

· TythonLee opened on Jun 9, 2025

可否提供一下dpo的pipeline

Status: Open.

#1359 In FunAudioLLM/CosyVoice;

· fantasyyyy opened on Jun 7, 2025

<code>KeyError: 'prompt_text_len'</code> (之前遇到的 <code>KeyError: 'prompt_text'</code>)

Status: Open.

#1358 In FunAudioLLM/CosyVoice;

· Edwinhei opened on Jun 6, 2025

使用指定音色id时报错llm_prompt_speech_token

Status: Open.

#1357 In FunAudioLLM/CosyVoice;

· goudemaoningsir opened on Jun 6, 2025

vllm用的是LLMEngine阻塞推理的模式，多并发的情况下表现不理想，有考虑使用非阻塞的AsyncLLMEngine吗

Status: Open.

#1356 In FunAudioLLM/CosyVoice;

· 0xCAFEBABE0 opened on Jun 6, 2025

0603新代码用vllm进行多并发推理时部分音频片段异常

Status: Open.

#1355 In FunAudioLLM/CosyVoice;

· WendongGan opened on Jun 6, 2025

vllm部署模型如何做到并发呢？

Status: Open.

#1354 In FunAudioLLM/CosyVoice;

· hhhhadiii opened on Jun 5, 2025

vllm加速后会出现吞字漏字的现象怎么解决

Status: Open.

#1353 In FunAudioLLM/CosyVoice;

· Chenruxin opened on Jun 5, 2025