LyAhn 98e2bf9237 perf: migrate to JezzWTF/VibeVoice fork, parallel CFG executors ...

Switch vibevoice dependency from microsoft/VibeVoice to JezzWTF/VibeVoice
fork (commit e76701f) which contains the async decode + parallel CFG
optimisations directly in generate(). Removes the instance-method
patching approach (vibevoice_generate_patch.py deleted).

server/vibevoice_server.py:
- Add _cfg_executor (ThreadPoolExecutor, 1 worker) alongside _decode_executor
- _install_cpu_pipeline_optimizations now sets both executors directly as
  model._vibepod_decode_executor and model._vibepod_cfg_executor
- Both executors shut down in lifespan on exit
- Remove vibevoice_generate_patch import/install (no longer needed)

server/pyproject.toml:
- vibevoice source changed to git+https://github.com/JezzWTF/VibeVoice.git
- No machine-local paths; works identically on any clone

2026-04-30 21:30:07 +01:00

..

.python-version

perf: improve streaming generation pipeline

2026-04-30 18:54:14 +01:00

download_model.py

Improve CPU Inference Stability: Adaptive Buffering & Chunk Accumulation (#11)

2026-04-30 16:03:35 +01:00

pyproject.toml

perf: migrate to JezzWTF/VibeVoice fork, parallel CFG executors

2026-04-30 21:30:07 +01:00

start.sh

perf: CPU async pipeline overlap + INT8 quantization

2026-04-30 20:46:29 +01:00

uv.lock

perf: migrate to JezzWTF/VibeVoice fork, parallel CFG executors

2026-04-30 21:30:07 +01:00

vibevoice_server.py

perf: migrate to JezzWTF/VibeVoice fork, parallel CFG executors

2026-04-30 21:30:07 +01:00