Output tokens >> reasoning tokens to avoid early response termination.

This commit is contained in:
Debanjum
2025-05-18 14:40:57 -07:00
parent 73e28666b5
commit 31dcc44c20

View File

@@ -36,7 +36,9 @@ logger = logging.getLogger(__name__)
gemini_clients: Dict[str, genai.Client] = {} gemini_clients: Dict[str, genai.Client] = {}
MAX_OUTPUT_TOKENS_GEMINI = 8192 # Output tokens should be more than reasoning tokens.
# This avoids premature response termination.
MAX_OUTPUT_TOKENS_GEMINI = 20000
MAX_REASONING_TOKENS_GEMINI = 10000 MAX_REASONING_TOKENS_GEMINI = 10000
SAFETY_SETTINGS = [ SAFETY_SETTINGS = [