Pinned Loading

1. PrimeMinutes PrimeMinutes Public

   Fully offline, on-device Android port of VoxSum Studio (transcribe + summarize audio).

   Swift 2

2. meeting-summarizer meeting-summarizer Public

   On-device long-meeting summarization (zh-TW): map-reduce distillation, transcript-grounded evaluation, short-context reading agent

   Python

3. nemo-x-asr-diarizer.cpp nemo-x-asr-diarizer.cpp Public

   Streaming x-asr + Nemotron-3 diarization in one process: 5x faster than a 1.5B ASR+diarize LLM at 1/5 the memory, with speaker attribution 30x more accurate on hard bilingual audio

   C++

4. PULSAR-ASM PULSAR-ASM Public

   Forked from tomtsai28/PULSAR-ASM

   Pure x86-64 assembly LLM engine for Gemma-2B in 5KB flat machine code. Zero runtimes, memory-saturating AVX2+F16C execution.

   Cuda

5. voice-chat voice-chat Public

   Python

6. sloth-zhuyin-linux sloth-zhuyin-linux Public

   懶音輸入法(樹懶注音輸入法/Slothing)— LLM-powered Zhuyin IME: 免選字整句轉換, libchewing-free. fcitx5 / IBus / Android / Web.

   C++ 3