diff options
| author |  sommerfeld <sommerfeld@sommerfeld.dev> | 2026-05-13 13:43:23 +0100 |
|---|---|---|
| committer | sommerfeld <sommerfeld@sommerfeld.dev> | 2026-05-13 13:43:23 +0100 |
| commit | d643aaf8a3a3fc8a31f39cb2ae3eaaf681300d4e (patch) | |
| tree | d3a030b6955093707d095fbe37a9c5ef1137ae93 /dot_local | |
| parent | 5c241d65ed4a6ec2bc3e5d75d6858ed6722f1b17 (diff) | |
| download | dotfiles-d643aaf8a3a3fc8a31f39cb2ae3eaaf681300d4e.tar.gz dotfiles-d643aaf8a3a3fc8a31f39cb2ae3eaaf681300d4e.tar.bz2 dotfiles-d643aaf8a3a3fc8a31f39cb2ae3eaaf681300d4e.zip | |
perf(dictate): switch default model to base for ~5x speedup
large-v3-turbo-q5_0 ran ~1-2x realtime on the T490's CPU, making
push-to-talk feel sluggish. The base multilingual model is ~142 MB
(vs 547 MB) and runs ~7-10x realtime, dropping perceived latency on
short utterances from a few seconds to near-instant.
Quality on short EN/PT dictation remains usable; bump WHISPER_MODEL
to small or large-v3-turbo if accuracy matters more than latency.
Diffstat (limited to 'dot_local')
| -rw-r--r-- | dot_local/bin/executable_dictate | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/dot_local/bin/executable_dictate b/dot_local/bin/executable_dictate index 2e8a361..ae8b14c 100644 --- a/dot_local/bin/executable_dictate +++ b/dot_local/bin/executable_dictate @@ -16,7 +16,7 @@ state_dir="${XDG_RUNTIME_DIR:-/tmp}/dictate" pid_file="$state_dir/pid" wav_file="$state_dir/audio.wav" log_file="$state_dir/whisper.log" -default_model='/usr/share/whisper.cpp-model-large-v3-turbo-q5_0/ggml-large-v3-turbo-q5_0.bin' +default_model='/usr/share/whisper.cpp-model-base/ggml-base.bin' model="${WHISPER_MODEL:-$default_model}" lang="${WHISPER_LANG:-auto}" @@ -34,7 +34,7 @@ start_recording() { fi if [ ! -r "$model" ]; then notify-send -u critical "🎙️ dictate" \ - "Model missing: $model. Install whisper.cpp-model-large-v3-turbo-q5_0 (AUR)." + "Model missing: $model. Install whisper.cpp-model-base (AUR)." exit 1 fi rm -f "$wav_file" |