llama-stack-mirror/CHANGELOG.md

Changelog

0.0.53

Added

• Resource-oriented design for models, shields, memory banks, datasets and eval tasks
• Persistence for registered objects with distribution
• Ability to persist memory banks created for FAISS
• PostgreSQL KVStore implementation
• Environment variable placeholder support in run.yaml files
• Provider deprecation support
• Comprehensive Zero-to-Hero notebooks and quickstart guides
• Colab notebook integration for Llama Stack lesson
• Remote::vllm provider with vision model support
• Support for quantized models in Ollama
• Vision models support for Together, Fireworks, Meta-Reference, and Ollama
• Bedrock distribution with safety shields support

Changed

• Split safety into distinct providers (llama-guard, prompt-guard, code-scanner)
• Improved distribution registry with SQLite default storage
• Enhanced test infrastructure with composable fixtures
• Changed provider naming convention (impls → inline, adapters → remote)
• Updated API signatures for dataset and eval task registration
• Restructured folder organization for providers
• Enhanced Docker build configuration
• Added version prefixing for REST API routes

Removed

• llama stack configure command

Fixed

• Agent system instructions persistence after first turn
• PostgreSQL config validation
• vLLM adapter chat completion signature
• Routing table lookup key for memory banks
• Together inference validator
• Exception handling for SSE connection closure

Development

• Added new test migration system
• Enhanced test setup with configurable model selection
• Added support for remote providers in tests
• Improved pre-commit hooks configuration
• Updated OpenAPI generator and specifications