Commit History
12 days ago
-
feat(portfolio): unify Llama-3.2-1B and Qwen2.5-0.5B models with dynamic switching and updated docs
51d622b
-
feat(wearos): add Samsung Galaxy Watch 4 streaming support, 14 bug fixes, and clinical benchmarks
- Add WearOSPPGAdapter with fast DC baseline stripping, 100 Hz -> 25 Hz anti-aliased FIR decimation, multi-parameter SQI, and rolling 90s ring buffer - Implement Wear OS test bench simulator and production Android Kotlin companion streaming architecture - Resolve 14 critical codebase bugs across threading, numerical stability, macOS APFS locks, and low-parameter generation - Enhance on-device Clinical RAG engine with condition-specific intent boosting (+30/-10) - Update 75-waveform biosignal benchmark (100% accuracy) and multi-domain clinical reasoning evaluation - Update full documentation in DOCUMENTATION.md and README.mdefc501a
-
fix(chat): resolve Q&A completion trigger, enhance RAG grounding, and add benchmark audit suite
- Fix RAG completion mimicry by reformatting retrieved guidelines from Q&A to neutral grounding blocks - Disentangle sensor telemetry directives from general lifestyle and medication inquiries - Eliminate chat history duplication and strip disclaimer repetition in conversational memory - Add fail-safe fallback to prevent empty disclaimer-only responses - Prioritize foundational ACC/AHA nutritional and pharmacotherapy guidelines - Add comprehensive benchmark accuracy test suite and update project documentation6198e11