Spaces:
Running
Running
File size: 1,236 Bytes
f06dea6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 | # # llm_engine.py
import google.generativeai as genai
from app.core.config import GEMINI_API_KEY
from langchain_google_genai import ChatGoogleGenerativeAI
# ✅ Configure Gemini client
genai.configure(api_key=GEMINI_API_KEY)
llm = ChatGoogleGenerativeAI(
model="gemini-2.5-flash",
google_api_key=GEMINI_API_KEY,
temperature=0.2,
max_output_tokens=800,
)
# # ✅ Separate LLM for evaluator — needs near-deterministic JSON output
# eval_llm = ChatGoogleGenerativeAI(
# model="gemini-2.5-flash",
# google_api_key=GEMINI_API_KEY,
# temperature=0.0, # ✅ deterministic — evaluator must return valid JSON
# max_output_tokens=200, # ✅ evaluator only returns a small JSON blob
# thinking_level="none" # to disable chain-of-thought
# )
eval_llm = ChatGoogleGenerativeAI(
model="gemini-2.0-flash", # no thinking, faster
google_api_key=GEMINI_API_KEY,
temperature=0.0,
max_output_tokens=200,
# model_kwargs={
# "generation_config": {
# "thinking_config": {
# "thinking_budget": 0 # ✅ 0 = disabled, bypasses langchain validation entirely
# }
# }
# }
)
|