Refa: remove temperature since some LLMs fail to support. (#8981)

### What problem does this PR solve? ### Type of change - [x] Refactoring
2025-12-08 20:42:30 +08:00 · 2025-07-23 10:17:04 +08:00
parent 0020c50000
commit 935ce872d8
7 changed files with 10 additions and 12 deletions
--- a/rag/raptor.py
+++ b/rag/raptor.py
@ -107,7 +107,7 @@ class RecursiveAbstractiveProcessing4TreeOrganizedRetrieval:
                            ),
                        }
                    ],
-                    {"temperature": 0.3, "max_tokens": self._max_token},
+                    {"max_tokens": self._max_token},
                )
                cnt = re.sub(
                    "(······\n由于长度的原因，回答被截断了，要继续吗？|For the content length reason, it stopped, continue?)",