Changelog

Tiny-LLM keeps a short, release-oriented changelog. This surface is reserved for meaningful external milestones.

Language Selection

Releases

[2.0.2] — 2026-04-27

Added

  • Quantization utilities: New quantization.h/quantization.cpp with F32→F16, Q4_0, Q8_0, W8A16 utilities
  • CLI enhancements: --help, --version, --info options for better usability

Changed

  • KVCacheManager: Factory method create() for consistent Result<T> error handling
  • Code quality: Added noexcept to simple accessors, fixed clang-format-18 violations

Infrastructure

  • CI simplification with Jimver/cuda-toolkit action
  • Enhanced .clangd LSP configuration
Details: English 中文

[2.0.1] — 2026-04-16

Fixed

  • CRITICAL: QuantizedWeight scale dimension calculation error in test utilities
  • Removed unused code in attention kernel (q_reg array loading)
Details: English 中文

[2.0.0] — 2026-03-09

Changed ⚠️ BREAKING

  • API Redesign: KVCache appendKV() is now stateless with explicit advanceSeqLen()
  • CMake modernization with target exports and architecture auto-detection

Added

  • CI workflow with automated format checking
  • tiny_llm::tiny_llm CMake alias target

Migration Guide: Update any direct KVCache usage to call advanceSeqLen() after all layers.

Details: English 中文

Notes

  • Infrastructure-only cleanup is intentionally excluded from this public changelog.
  • For the current project story and onboarding flow, start at Home or Documentation.
← Home 中文更新日志 Contributing

Back to top