BGE-M3 Embedded Model

BGE-M3 is the vector embedding service built into the AI Pod, suitable for semantic retrieval, RAG recall, and text similarity tasks.

Service Information

• Service address:
• Embeddings address:
• Default model name: bge-m3
• Container image: registry.lazycat.cloud/x/lzc-aipod-bge-m3:a7d1e2b

Standalone Deployment

1. Save the following content into a docker-compose.yml file.

services:
  bgem3vllm:
    image: registry.lazycat.cloud/x/lzc-aipod-bge-m3:a7d1e2b
    ports:
      - 8000:8000
    environment:
      - LANGUAGE=en_US:en
      - LANG=en_US.UTF-8
      - LC_ALL=en_US.UTF-8
      - VLLM_CONFIGURE_LOGGING=0
      - VLLM_DO_NOT_TRACK=1
      - VLLM_NO_USAGE_STATS=1
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 10s
      timeout: 30s
      retries: 100
      start_period: 360s

2. Run docker-compose up -d in the current directory to start
3. After startup, you can call http://127.0.0.1:8000/v1/embeddings to invoke

Invocation Examples

curl http://127.0.0.1:8000/v1/embeddings \
  -H "Content-Type: application/json" \
  -d '{"model":"bge-m3","input":"算力舱文档检索测试"}'