GeneralBots/botserver
2
1
Fork 0
Code Activity Actions 4

fix: GLM max_tokens 131072, disable thinking kwargs - model sends content after reasoning naturally
All checks were successful
BotServer CI/CD / build (push) Successful in 3m11s
Details

Browse source
This commit is contained in:
Rodrigo Rodriguez (Pragmatismo) 2026-04-13 18:52:02 -03:00
parent c9fa057203
commit d6ffe265ef
1 changed files with 4 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
src/llm/glm.rs
View file

@ -153,15 +153,12 @@ impl LLMProvider for GLMClient {
model: model_name.to_string(),
messages,
stream: Some(false),
max_tokens: Some(16384),
max_tokens: Some(131072),
temperature: Some(1.0),
top_p: Some(1.0),
tools: None,
tool_choice: None,
chat_template_kwargs: Some(GLMChatTemplateKwargs {
enable_thinking: true,
clear_thinking: false,
}),
chat_template_kwargs: None,
};
let url = self.build_url();
@ -242,15 +239,12 @@ impl LLMProvider for GLMClient {
model: model_name.to_string(),
messages,
stream: Some(true),
max_tokens: Some(16384),
max_tokens: Some(131072),
temperature: Some(1.0),
top_p: Some(1.0),
tools: tools.cloned(),
tool_choice,
chat_template_kwargs: Some(GLMChatTemplateKwargs {
enable_thinking: true,
clear_thinking: false,
}),
chat_template_kwargs: None,
};
let url = self.build_url();