// Run performance benchmarks. Arguments: asr, tts, vad, diarize (with optional --num-files N).
Build the speech-swift package (release or debug). Use when preparing for testing, benchmarking, or running demos.
Run tests. Use after code changes to validate. Arguments: unit (default, no GPU), e2e (with models), filter name, or all.
| name | benchmark |
| description | Run performance benchmarks. Arguments: asr, tts, vad, diarize (with optional --num-files N). |
Run benchmarks using the release build. Build first with /build.
/benchmark asr — transcribe test audio, report RTF/benchmark tts — synthesize test text, report RTF/benchmark vad — VAD on VoxConverse (all engines)/benchmark diarize — DER on VoxConverse (requires downloaded test set)module="$ARGUMENTS"
cli=".build/release/speech"
case "$module" in
asr)
$cli transcribe Tests/Qwen3ASRTests/Resources/test_audio.wav 2>&1
;;
tts)
$cli speak "The quick brown fox jumps over the lazy dog." --output /tmp/bench_tts.wav 2>&1
;;
vad)
python3 scripts/benchmark_vad.py --compare --num-files 5 2>&1
;;
diarize)
python3 scripts/benchmark_diarization.py --num-files 5 2>&1
;;
*)
echo "Usage: /benchmark [asr|tts|vad|diarize]"
;;
esac
| Module | Metric | Target |
|---|---|---|
| ASR (Qwen3) | RTF | ~0.06 |
| ASR (Parakeet) | RTF | ~0.025 |
| TTS | RTF | ~0.7 |
| VAD (Silero) | RTF | >20x real-time |
| Diarization | DER | <10% (VoxConverse) |