klbr/khoj
1
0
Fork 0
mirror of https://github.com/khoaliber/khoj.git synced 2026-03-04 13:20:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Reduce max reasoning tokens for gemini models

Browse Source 
A high reasoning tokens does not seem to help for standard Khoj use
cases. And hopefully reducing it may avoid repetition loops by model.
This commit is contained in:
Debanjum
2025-08-22 18:57:38 -07:00
parent 2823c84bb4
commit 8aa9c0f534
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/khoj/processor/conversation/google/utils.py
View File

@@ -46,7 +46,7 @@ gemini_clients: Dict[str, genai.Client] = {}
# This avoids premature response termination.
MAX_OUTPUT_TOKENS_FOR_REASONING_GEMINI = 20000
MAX_OUTPUT_TOKENS_FOR_STANDARD_GEMINI = 8000
MAX_REASONING_TOKENS_GEMINI = 10000
MAX_REASONING_TOKENS_GEMINI = 512
SAFETY_SETTINGS = [
gtypes.SafetySetting(