feat: integrate RouteLLM BERT router for intelligent query classification
- 添加 transformers 和 torch 依赖 - 创建 bert_router.py 封装 RouteLLM BERT 分类器 - 新增 select_model_by_bert() 函数替代 token 长度路由 - BERT 输出映射: strong->qwen-max, weak->qwen-flash - 保留 token 长度路由作为 fallback
This commit is contained in:
@@ -5,5 +5,7 @@ litellm>=1.0.0
|
||||
tiktoken>=0.5.0
|
||||
httpx>=0.25.0
|
||||
python-dotenv>=1.0.0
|
||||
transformers>=4.30.0
|
||||
torch>=2.0.0
|
||||
pytest>=7.4.0
|
||||
pytest-asyncio>=0.21.0
|
||||
|
||||
Reference in New Issue
Block a user