vrr/ruvector
2
0
Fork 0
mirror of https://github.com/ruvnet/RuVector.git synced 2026-05-25 23:24:03 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
ruvector/docker/images/ruvector-sona
History
Exact
rUv aca7c53bc8 feat(docker): Add 8 specialized Docker images with publishing infrastructure 
- Add Dockerfiles for 8 RuVector components:
  - ruvector-core: Core vector database engine with HNSW indexing
  - ruvector-server: REST API server (port 8080)
  - ruvector-cli: CLI + MCP server for AI integration (port 3000)
  - ruvector-gnn: Graph Neural Networks (GCN, GAT, GIN)
  - ruvector-graph: Neo4j-compatible Cypher graph DB (ports 7687, 7474)
  - ruvector-attention: 39 attention mechanisms (MHA, GQA, MoA)
  - ruvector-cluster: Raft consensus distributed clustering
  - ruvector-sona: Self-Optimizing Neural Architecture

- Add comprehensive README.md for each image with:
  - Docker Hub badges
  - Features and quickstart guides
  - Configuration tables
  - Performance benchmarks

- Add docker-compose.full.yml for 9-service orchestration
- Add build/publish/test scripts in docker/scripts/
- Add GitHub Actions workflow for multi-arch Docker publishing

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-07 04:24:28 +00:00
..
Dockerfile feat(docker): Add 8 specialized Docker images with publishing infrastructure 2025-12-07 04:24:28 +00:00
README.md feat(docker): Add 8 specialized Docker images with publishing infrastructure 2025-12-07 04:24:28 +00:00

README.md

RuVector SONA

Docker Pulls Docker Image Size Docker Image Version GitHub

Self-Optimizing Neural Architecture for runtime-adaptive learning. SONA enables continuous learning without retraining using LoRA, EWC++, and ReasoningBank.

Features

  • 🧠 Self-learning - Improves results over time
  • 🔧 Two-tier LoRA - Efficient low-rank adaptation
  • 🛡️ EWC++ - Prevents catastrophic forgetting
  • 📚 ReasoningBank - Pattern storage and retrieval
  • 📈 10-30% accuracy improvement

Quick Start

docker run -d --name ruvector-sona -p 8085:8085 ruvnet/ruvector-sona:latest

Learning API

# Enable learning
curl -X POST http://localhost:8085/learning/enable -d '{"enabled": true}'

# Record feedback
curl -X POST http://localhost:8085/feedback -d '{"query_id": "q1", "relevance": 0.95}'

# Auto-tune
curl -X POST http://localhost:8085/autotune

Configuration

Variable Default Description
SONA_PORT 8085 Service port
LEARNING_RATE 0.001 Learning rate
LORA_RANK 16 LoRA rank
EWC_LAMBDA 0.4 EWC strength

Performance

Samples Improvement
1,000 +15%
10,000 +25%
100,000 +30%

License

MIT License