apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 16:48:03 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

Support limit thinking lengths (#4070)

Browse Source 
Co-authored-by: K11OntheBoat <“ruianmaidanglao@163.com”>
This commit is contained in:
K11OntheBoat
2025-09-17 12:40:08 +08:00
committed by GitHub
parent b41988f4bc
commit 7f9a9b37f3
8 changed files with 184 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
fastdeploy/entrypoints/engine_client.py
View File

@@ -177,8 +177,6 @@ class EngineClient:
task["prompt_token_ids_len"] = len(task["prompt_token_ids"])
input_ids_len = task["prompt_token_ids_len"]
task["max_tokens"] = min(self.max_model_len - input_ids_len, task.get("max_tokens"))
if task.get("reasoning_max_tokens", None) is None:
task["reasoning_max_tokens"] = max(int(task["max_tokens"] * 0.8), 1)
min_tokens = task.get("min_tokens", 1)
if "messages" in task:
del task["messages"]