deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-04 09:29:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

update

Browse Source
This commit is contained in:
lyuxiang.lx
2025-04-07 22:45:47 +08:00
parent d9ffd592f6
commit 4951d2ad1a
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
cosyvoice/cli/model.py
View File

@@ -355,6 +355,7 @@ class CosyVoice2Model(CosyVoiceModel):
opt_shape = [(2, 80, 200), (2, 1, 200), (2, 80, 200), (2, 80, 200), (1, 4, 2, 100, 512, 2), (12, 4, 2, 100, 512, 2), (1, 4, 2, 100, 512, 2)]
max_shape = [(2, 80, 1500), (2, 1, 1500), (2, 80, 1500), (2, 80, 1500), (1, 4, 2, 200, 512, 2), (12, 4, 2, 200, 512, 2), (1, 4, 2, 200, 512, 2)]
input_names = ["x", "mask", "mu", "cond", 'down_blocks_kv_cache', 'mid_blocks_kv_cache', 'up_blocks_kv_cache']
assert self.use_flow_cache is True, "get_trt_kwargs is set for flow cache mode. If you want to use trt with use_flow_cache=False, please set higher max_shape"
return {'min_shape': min_shape, 'opt_shape': opt_shape, 'max_shape': max_shape, 'input_names': input_names}
def token2wav(self, token, prompt_token, prompt_feat, embedding, uuid, finalize=False, speed=1.0):