cnphpbb/deploy.stack
1
0
Fork 0
forked from DevOps/deploy.stack
Code Pull Requests Activity

feat(honcho): embedding 切换 SiliconFlow Qwen3-0.6B 支持 Ollama 备选

Browse Source 
- embedding 默认走 SiliconFlow Qwen3-Embedding-0.6B（独立 base URL）
- env.cfg.example 加 Ollama 备选注释块（host.docker.internal:11434）
- stack.yml api+deriver 双服务 EMBEDDING_* 配置同步
- api 段 EMBEDDING_API_KEY 用 :- 默认空，兼容 Ollama 无 key 场景
- readme Mermaid 架构图加 siliconflow + ollama 节点
- HONCHO_DERIVER_MODEL 默认值 gpt-4o-mini → gpt-5.4-mini（与上游对齐）
This commit is contained in:
cnphpbb
2026-06-19 15:12:50 +08:00
parent 48b52ca21c
commit 9c08233d4e
3 changed files with 40 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
honcho/stack.yml
+10 -2
View File
@@ -42,7 +42,10 @@ services:
- NAMESPACE=${HONCHO_NAMESPACE:-honcho}
- EMBED_MESSAGES=${HONCHO_EMBED_MESSAGES:-true}
- EMBEDDING_MODEL_CONFIG__TRANSPORT=openai
- EMBEDDING_MODEL_CONFIG__MODEL=${HONCHO_EMBEDDING_MODEL:-text-embedding-3-small}
- EMBEDDING_MODEL_CONFIG__MODEL=${HONCHO_EMBEDDING_MODEL:-Qwen/Qwen3-Embedding-0.6B}
- EMBEDDING_MODEL_CONFIG__OVERRIDES__BASE_URL=${HONCHO_EMBEDDING_BASE_URL:-https://api.siliconflow.cn/v1}
- EMBEDDING_VECTOR_DIMENSIONS=1024
- EMBEDDING_API_KEY=${HONCHO_EMBEDDING_API_KEY:-}
- DERIVER_MODEL_CONFIG__TRANSPORT=openai
- DERIVER_MODEL_CONFIG__MODEL=${HONCHO_DERIVER_MODEL:-gpt-4o-mini}
- DERIVER_WORKERS=${HONCHO_DERIVER_WORKERS:-1}
@@ -88,9 +91,14 @@ services:
- CACHE_ENABLED=true
- NAMESPACE=${HONCHO_NAMESPACE:-honcho}
- DERIVER_MODEL_CONFIG__TRANSPORT=openai
- DERIVER_MODEL_CONFIG__MODEL=${HONCHO_DERIVER_MODEL:-gpt-4o-mini}
- DERIVER_MODEL_CONFIG__MODEL=${HONCHO_DERIVER_MODEL:-gpt-5.4-mini}
- DERIVER_WORKERS=${HONCHO_DERIVER_WORKERS:-1}
- EMBEDDING_MODEL_CONFIG__TRANSPORT=openai
- EMBEDDING_MODEL_CONFIG__MODEL=${HONCHO_EMBEDDING_MODEL:-Qwen/Qwen3-Embedding-0.6B}
- EMBEDDING_MODEL_CONFIG__OVERRIDES__BASE_URL=${HONCHO_EMBEDDING_BASE_URL:-https://api.siliconflow.cn/v1}
- EMBEDDING_VECTOR_DIMENSIONS=1024
- LLM_OPENAI_API_KEY=${HONCHO_LLM_API_KEY:?set HONCHO_LLM_API_KEY}
- EMBEDDING_API_KEY=${HONCHO_EMBEDDING_API_KEY:-}
- MODEL_CONFIG__OVERRIDES__BASE_URL=${HONCHO_LLM_BASE_URL:-https://api.openai.com/v1}
volumes:
- /etc/timezone:/etc/timezone:ro