AXERA-TECH
/

CosyVoice3

Model card Files Files and versions

lihongjie commited on 29 days ago

Commit

6ca7c7e

·

1 Parent(s): ca08716

update

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -35,8 +35,8 @@ For those who are interested in model conversion, you can try to export axmodel
 | Stage | Time |
 |------|------|
 | llm prefill ( input_token_num + prompt_token_num 在 [0,128 ] ) | 104 ms  |
-| llm prefill ( input_token_num + prompt_token_num 在 [128,256 ] ) | 234 ms  |
-| Decode  |  21.24 token/s |
 ## How to use

 | Stage | Time |
 |------|------|
 | llm prefill ( input_token_num + prompt_token_num 在 [0,128 ] ) | 104 ms  |
+| llm prefill ( input_token_num + prompt_token_num 在 [128,256 ] ) | 160 ms  |
+| Decode  |  14 token/s |
 ## How to use