[Feature] add a new reasoning parser #4571
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thanks for your contribution! |
… reasoning_parser
LiqinruiG
reviewed
Oct 28, 2025
| if self.think_end_token not in model_output: | ||
| if self.tool_begin_token not in model_output: | ||
| return "", model_output | ||
| else: |
Collaborator
There was a problem hiding this comment.
上面的 if 已经 return 了， 这里不需要 else。
| if self.tool_call_start_token_id not in current_token_ids: | ||
| return DeltaMessage(content=delta_text) | ||
| # Skip empty chunks | ||
| if len(delta_text.strip()) == 0: |
| # Process the buffer content | ||
| if "<tool_call>" in delta_text: | ||
| if self.valid is None: | ||
| tool_call_begin = self.buffer.find(self.tool_call_start_token) |
Collaborator
There was a problem hiding this comment.
使用 current_texts， 不用buff
| return reasoning_content, "" | ||
| return reasoning_content, content | ||
|
|
||
| def _is_with_tool(self, current_text: str, current_token_ids: Sequence[int]) -> bool: |
| # Skip single special tokens | ||
| if len(delta_token_ids) == 1 and delta_token_ids[0] == self.think_end_token_id: | ||
| return None | ||
| if not self._is_with_tool(current_text=current_text, current_token_ids=current_token_ids): |
Collaborator
There was a problem hiding this comment.
- 不存在
- with_tool = true
2.1 <tool_call> in delta: 取前面内容
2.2 None - with_tool = false
3.1 in delta
后面如果都是\n， 回复内容为空， 否则区分思考和回复
3.2 current 判断 <\think> 后面是否都是 \n。
kxz2002
added a commit
to kxz2002/FastDeploy
that referenced
this pull request
Oct 29, 2025
* add new reasoning_parser initial commit * add parser file content * add register * ernie_test_reasoning_parser * support <tool_call> token and add tool_parser * add and fix unit tests * modify reasoning_parser * modify reasoning parser and tool parser * modify unit tests * modify reasoning_parser and tool_parser * modify unit tests * fix tool_parser * modify the logic of reasoning_parser and tool_parser * add and modify unit tests * standardize code style * simplify reasoning_parser and tool_parser * modify unit test
Jiang-Jia-Jun
pushed a commit
that referenced
this pull request
Oct 30, 2025
* add new reasoning_parser initial commit * add parser file content * add register * ernie_test_reasoning_parser * support <tool_call> token and add tool_parser * add and fix unit tests * modify reasoning_parser * modify reasoning parser and tool parser * modify unit tests * modify reasoning_parser and tool_parser * modify unit tests * fix tool_parser * modify the logic of reasoning_parser and tool_parser * add and modify unit tests * standardize code style * simplify reasoning_parser and tool_parser * modify unit test
Jiang-Jia-Jun
pushed a commit
that referenced
this pull request
Oct 31, 2025
… and reasoning_parser to “-” (#4668) (#4737) * [Feature] add a new reasoning parser (#4571) * add new reasoning_parser initial commit * add parser file content * add register * ernie_test_reasoning_parser * support <tool_call> token and add tool_parser * add and fix unit tests * modify reasoning_parser * modify reasoning parser and tool parser * modify unit tests * modify reasoning_parser and tool_parser * modify unit tests * fix tool_parser * modify the logic of reasoning_parser and tool_parser * add and modify unit tests * standardize code style * simplify reasoning_parser and tool_parser * modify unit test * [BugFix] Fix finish reason in _create_chat_completion_choice (#4582) * fix n_param _create_chat_completion_choicel * fix unit test * fix final_res * modify unit tests * [BugFix] fix offline llm chat "enable_thinking" is always "False" (#4686) * fix enable_thinking * recover ernie4_5_vl_processor * [Feature] Unify the registration name recognition for tool_parser and reasoning_parser to “-” (#4668) * parser register name unify * change ernie_x1 to ernie-x1 * change ernie4_5_vl to ernie-45-vl * fix unit test
Jiang-Jia-Jun
pushed a commit
that referenced
this pull request
Nov 1, 2025
) (#4716) * [Feature] add a new reasoning parser (#4571) * add new reasoning_parser initial commit * add parser file content * add register * ernie_test_reasoning_parser * support <tool_call> token and add tool_parser * add and fix unit tests * modify reasoning_parser * modify reasoning parser and tool parser * modify unit tests * modify reasoning_parser and tool_parser * modify unit tests * fix tool_parser * modify the logic of reasoning_parser and tool_parser * add and modify unit tests * standardize code style * simplify reasoning_parser and tool_parser * modify unit test * [BugFix] Fix finish reason in _create_chat_completion_choice (#4582) * fix n_param _create_chat_completion_choicel * fix unit test * fix final_res * modify unit tests * [BugFix] fix offline llm chat "enable_thinking" is always "False" (#4686) * fix enable_thinking * recover ernie4_5_vl_processor
Jiang-Jia-Jun
pushed a commit
that referenced
this pull request
Nov 5, 2025
…an not disable thinking (#4762) (#4798) * [Feature] add a new reasoning parser (#4571) * add new reasoning_parser initial commit * add parser file content * add register * ernie_test_reasoning_parser * support <tool_call> token and add tool_parser * add and fix unit tests * modify reasoning_parser * modify reasoning parser and tool parser * modify unit tests * modify reasoning_parser and tool_parser * modify unit tests * fix tool_parser * modify the logic of reasoning_parser and tool_parser * add and modify unit tests * standardize code style * simplify reasoning_parser and tool_parser * modify unit test * [BugFix] Fix finish reason in _create_chat_completion_choice (#4582) * fix n_param _create_chat_completion_choicel * fix unit test * fix final_res * modify unit tests * [BugFix] fix offline llm chat "enable_thinking" is always "False" (#4686) * fix enable_thinking * recover ernie4_5_vl_processor * [BugFix] Fix ernie4_5_vl_processor.py and qwen_vl_processor.py can not disable thinking (#4762) * fix ernie4_5_vl_processor.py and qwen_vl_processor.py * add unit test
Jiang-Jia-Jun
pushed a commit
that referenced
this pull request
Nov 6, 2025
* [Feature] add a new reasoning parser (#4571) * add new reasoning_parser initial commit * add parser file content * add register * ernie_test_reasoning_parser * support <tool_call> token and add tool_parser * add and fix unit tests * modify reasoning_parser * modify reasoning parser and tool parser * modify unit tests * modify reasoning_parser and tool_parser * modify unit tests * fix tool_parser * modify the logic of reasoning_parser and tool_parser * add and modify unit tests * standardize code style * simplify reasoning_parser and tool_parser * modify unit test * [BugFix] Fix finish reason in _create_chat_completion_choice (#4582) * fix n_param _create_chat_completion_choicel * fix unit test * fix final_res * modify unit tests * [BugFix] fix offline llm chat "enable_thinking" is always "False" (#4686) * fix enable_thinking * recover ernie4_5_vl_processor * fix parser register name (#4795) Co-authored-by: luukunn <[email protected]> --------- Co-authored-by: luukunn <[email protected]>
Jiang-Jia-Jun
pushed a commit
that referenced
this pull request
Nov 6, 2025
…_end_token' (#4805) (#4842) * [Feature] add a new reasoning parser (#4571) * add new reasoning_parser initial commit * add parser file content * add register * ernie_test_reasoning_parser * support <tool_call> token and add tool_parser * add and fix unit tests * modify reasoning_parser * modify reasoning parser and tool parser * modify unit tests * modify reasoning_parser and tool_parser * modify unit tests * fix tool_parser * modify the logic of reasoning_parser and tool_parser * add and modify unit tests * standardize code style * simplify reasoning_parser and tool_parser * modify unit test * [BugFix] Fix finish reason in _create_chat_completion_choice (#4582) * fix n_param _create_chat_completion_choicel * fix unit test * fix final_res * modify unit tests * [BugFix] fix offline llm chat "enable_thinking" is always "False" (#4686) * fix enable_thinking * recover ernie4_5_vl_processor * [BugFix] Fix ernie_vl_reasoning_parsers.py 'end_token' to 'think_end_token' (#4805) * fix ernie_vl_reasoning_parsers.py 'end_token' to 'think_end_token' * add unit tests
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Motivation
Add a new reasoning_parser
The model's output format with reasoning includes two types:
No tool invocation: xxxyyy
Tool invocation: xxx\n<tool_call>zzz</tool_cal>
x represents the reasoning content, y represents the response content, and z represents the tool invocation content.
Modifications
Add a new reasoning_parser ErnieTestReasoningParser
Usage or Command
Accuracy Tests
Checklist
[FDConfig],[APIServer],[Engine],[Scheduler],[PD Disaggregation],[Executor],[Graph Optimization],[Speculative Decoding],[RL],[Models],[Quantization],[Loader],[OP],[KVCache],[DataProcessor],[BugFix],[Docs],[CI],[Optimization],[Feature],[Benchmark],[Others],[XPU],[HPU],[GCU],[DCU],[Iluvatar],[Metax]]pre-commitbefore commit.releasebranch, make sure the PR has been submitted to thedevelopbranch, then cherry-pick it to thereleasebranch with the[Cherry-Pick]PR tag.