ivada-infra/DOCS
3
0
Fork 0
Code Issues 1 Pull Requests 5 Actions Packages Projects Releases Wiki Activity

docs: update LLM model configuration to use DEFAULT_LLM_MODEL env var

Browse Source 
- Add DEFAULT_LLM_MODEL to environment variables section
- Update code examples to use settings.DEFAULT_LLM_MODEL instead of hardcoded model name
- Aligns with code refactoring to remove hardcoded model names
This commit is contained in:
Claude-51124 2025-11-18 20:16:17 +09:00
parent 42ee6f9cba
commit 9ad5468f9f
1 changed files with 8 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
journey/ideas/250804_LLM_모델_비교_분석.md
View File

@ -99,13 +99,20 @@ OPENAI_API_KEY=(OpenAI API 키)
MISTRAL_API_KEY=(Mistral API 키) MISTRAL_API_KEY=(Mistral API 키)
ANTHROPIC_API_KEY=(Anthropic API 키) ANTHROPIC_API_KEY=(Anthropic API 키)
# XAI_API_KEY=(xAI API 키) # 비용 문제로 제외 # XAI_API_KEY=(xAI API 키) # 비용 문제로 제외
# 기본 LLM 모델 설정 (모든 서비스에서 사용)
DEFAULT_LLM_MODEL=gemini-2.5-flash-lite
``` ```
### 사용 예시 ### 사용 예시
```python ```python
# 환경변수에서 기본 모델 가져오기 (권장)
from app.core.config import settings
default_model = settings.DEFAULT_LLM_MODEL # "gemini-2.5-flash-lite"
# 모델 선택 로직 # 모델 선택 로직
if task_type == "general": if task_type == "general":
model = "gemini-2.5-flash-lite" # 기본 model = settings.DEFAULT_LLM_MODEL # 기본 모델 사용
elif task_type == "quality": elif task_type == "quality":
model = "mistral-small-latest" # 품질 우선 model = "mistral-small-latest" # 품질 우선
elif task_type == "fast": elif task_type == "fast":