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

[fix]Modify follow-up push parameters and Modify the verification method for thinking length (#4086)

Browse Source 
* 续推参数  generated_token_ids 修改成 completion_token_ids;修改思考长度校验方式

* 续推参数  generated_token_ids 修改成 completion_token_ids;修改思考长度校验方式

* 续推参数  generated_token_ids 修改成 completion_token_ids;修改思考长度校验方式

* 续推参数  generated_token_ids 修改成 completion_token_ids;修改思考长度校验方式

* add completion_token_ids

* add logger

* fix reasoning_max_tokens ParameterError

* add unittest

* add unittest

* add unittest

* add unittest

* add unittest

* add unit test
This commit is contained in:
luukunn
2025-09-19 14:26:01 +08:00
committed by GitHub
parent 66a98b44ed
commit ee9d8a840a
6 changed files with 75 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
fastdeploy/entrypoints/engine_client.py
View File

@@ -236,8 +236,13 @@ class EngineClient:
raise ParameterError("max_tokens", f"max_tokens can be defined [1, {self.max_model_len}).")
if data.get("reasoning_max_tokens") is not None:
if data["reasoning_max_tokens"] > data["max_tokens"] or data["reasoning_max_tokens"] < 1:
raise ParameterError("reasoning_max_tokens", "reasoning_max_tokens must be between max_tokens and 1")
if data["reasoning_max_tokens"] < 1:
raise ParameterError("reasoning_max_tokens", "reasoning_max_tokens must be greater than 1")
if data["reasoning_max_tokens"] > data["max_tokens"]:
data["reasoning_max_tokens"] = data["max_tokens"]
api_server_logger.warning(
f"req_id: {data['request_id']}, reasoning_max_tokens exceeds max_tokens, the value of reasoning_max_tokens will be adjusted to match that of max_tokens"
)
# logprobs
logprobs = data.get("logprobs")