- I am using the OpenAI API to be the basic API spec.
- 我参照了OpenAI API作为基本的API。
- I am using the Result of
QSetting
as the configuration file. So, If you are using theServerGemma
only, you can generate a configuration file first by the QT version or just wrote it manually. Like below: - 我使用了 QSetting 的结果作为配置文件。
如果你打算直接使用
ServerGemma
,可以考虑使用QT
版本的gemma
先生成一次配置 文件,或者直接手写一份。参照下面示例:
[Setting]
MaxGeneratedTokens=2048
MaxTokens=3072
Temperature=1.0
Verbosity=2
ModelType=2b-pt
Weight=/opt/gemma/CodeGemma/2b-pt-sfp.sbs
Tokenizer=/opt/gemma/CodeGemma/tokenizer.spm
WebSocketPort=9999