Losita/mai-bot
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: 添加模型级别最大token数配置，并更新相关逻辑以支持优先级处理

Browse Source
This commit is contained in:
墨梓柒
2025-12-03 11:45:15 +08:00
parent 6680afaa4a
commit 12bc661790
3 changed files with 20 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/llm_models/utils_model.py
View File

@@ -324,11 +324,20 @@ class LLMRequest:
if effective_temperature is None:
effective_temperature = self.model_for_task.temperature
# max_tokens 优先级：参数传入 > 模型级别配置 > extra_params > 任务配置
effective_max_tokens = max_tokens
if effective_max_tokens is None:
effective_max_tokens = model_info.max_tokens
if effective_max_tokens is None:
effective_max_tokens = (model_info.extra_params or {}).get("max_tokens")
if effective_max_tokens is None:
effective_max_tokens = self.model_for_task.max_tokens
return await client.get_response(
model_info=model_info,
message_list=(compressed_messages or message_list),
tool_options=tool_options,
max_tokens=self.model_for_task.max_tokens if max_tokens is None else max_tokens,
max_tokens=effective_max_tokens,
temperature=effective_temperature,
response_format=response_format,
stream_response_handler=stream_response_handler,