iamdoubz/WhispAssist
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 13 Wiki Activity
766 Commits 1 Branch 13 Tags
31bdf0e39c2a2b0ee53ed626ac308ff20f18e8e5
Commit Graph
766 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
iamdoubz 31bdf0e39c fix(i18n): ru backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:34 -05:00
iamdoubz 5731e2df68 fix(i18n): pt-BR backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:33 -05:00
iamdoubz 04244196bb fix(i18n): pt-PT backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:32 -05:00
iamdoubz 28b9816162 fix(i18n): fr-CA backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:31 -05:00
iamdoubz a7d38b39dc fix(i18n): fr-FR backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:30 -05:00
iamdoubz 804b0a8598 fix(i18n): es-MX backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:29 -05:00
iamdoubz 804e3b6fa6 fix(i18n): es-ES backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:28 -05:00
iamdoubz 1525645ace fix(i18n): de backfill 53 untranslated keys, drop dead import.requires_1-3 2026-07-16 21:24:27 -05:00
iamdoubz d49f375a13 feat(i18n): ur strings for model quantization picker 2026-07-16 21:10:14 -05:00
iamdoubz 0f8b5500c2 feat(i18n): ar strings for model quantization picker 2026-07-16 21:10:13 -05:00
iamdoubz 9c349bb84d feat(i18n): si strings for model quantization picker 2026-07-16 21:10:12 -05:00
iamdoubz 5d2aa9840a feat(i18n): ta strings for model quantization picker 2026-07-16 21:10:11 -05:00
iamdoubz 204a507de7 feat(i18n): bn strings for model quantization picker 2026-07-16 21:10:10 -05:00
iamdoubz 4e167082ac feat(i18n): hi strings for model quantization picker 2026-07-16 21:10:08 -05:00
iamdoubz ee5c0ea7bc feat(i18n): ko strings for model quantization picker 2026-07-16 21:10:07 -05:00
iamdoubz 9311909a58 feat(i18n): zh-CN strings for model quantization picker 2026-07-16 21:10:06 -05:00
iamdoubz ec42d89f66 feat(i18n): fi strings for model quantization picker 2026-07-16 21:10:05 -05:00
iamdoubz c00da29f06 feat(i18n): nb strings for model quantization picker 2026-07-16 21:10:03 -05:00
iamdoubz 66b35c3a6c feat(i18n): sv strings for model quantization picker 2026-07-16 21:10:02 -05:00
iamdoubz 1a27fa0a1a feat(i18n): ru strings for model quantization picker 2026-07-16 21:10:01 -05:00
iamdoubz 8998863bb0 feat(i18n): pt-BR strings for model quantization picker 2026-07-16 21:09:59 -05:00
iamdoubz 94443dcd35 feat(i18n): pt-PT strings for model quantization picker 2026-07-16 21:09:58 -05:00
iamdoubz d3f4aeef14 feat(i18n): fr-CA strings for model quantization picker 2026-07-16 21:09:57 -05:00
iamdoubz 4bbaed7b0c feat(i18n): fr-FR strings for model quantization picker 2026-07-16 21:09:56 -05:00
iamdoubz b55bd5175f feat(i18n): es-MX strings for model quantization picker 2026-07-16 21:09:55 -05:00
iamdoubz d1b951e9ea feat(i18n): es-ES strings for model quantization picker 2026-07-16 21:09:54 -05:00
iamdoubz e2ec141d75 feat(i18n): de strings for model quantization picker 2026-07-16 21:09:53 -05:00
iamdoubz 90c12cb070 feat(i18n): en strings for model quantization picker 2026-07-16 21:09:51 -05:00
iamdoubz 88adce152b feat(ui): group whisper models per family with quantization dropdown (FR-MODEL-1) 2026-07-16 21:09:50 -05:00
iamdoubz 9e3f6327b7 feat(models): add Q8_0 variants + large-v3-turbo q5_0/q8_0 to catalog (FR-MODEL-1) 2026-07-16 21:06:35 -05:00
iamdoubz 8cdd0f61df Merge pull request 'Parakeet v Whisper h2h' (#26) from test_parakeet into main 
Reviewed-on: #26
v0.7.4 		2026-07-16 10:23:17 -05:00
iamdoubz 2d1c9fa478 Bench test npu 2026-07-16 10:10:30 -05:00
iamdoubz a71299803f docs(release): v0.7.4 release notes with SHA-256 checksums 2026-07-16 10:03:23 -05:00
iamdoubz fcf72da9db chore: bump version 0.7.4 2026-07-16 09:56:22 -05:00
iamdoubz 616d2f4f97 chore(bench): rustfmt asr_bench + allow large_enum_variant (one Engine per process) 2026-07-16 08:47:11 -05:00
iamdoubz be49c53940 docs(bench): W1 anomaly resolved — LP-E-core spin-barrier root cause, fixed numbers, criteria updated 2026-07-16 08:44:17 -05:00
iamdoubz 6cd736bec1 test(bench): W1 re-run after thread-cap fix — RTF 0.218, 0.827 cpu-sec/audio-sec, WER unchanged 2026-07-16 08:43:26 -05:00
iamdoubz 0d2eee7416 fix(transcription): cap whisper n_threads at 4 — hybrid-CPU spin-barrier pathology 
available_parallelism() put ggml workers on Meteor Lake LP E-cores; every
per-node spin barrier then waits on the slowest worker while the rest burn
CPU: RTF ~22 at 14 threads vs 0.28 at 4 threads (~80x), same build. This
was the real W1 CPU-pathology root cause; /Od-vs-/O2 made no difference.
4 matches whisper.cpp upstream default and is the CPU-frugal choice
(NFR-RES-1). WA_WHISPER_THREADS env overrides for support/bench.
 2026-07-16 08:39:45 -05:00
iamdoubz c7899e72a2 fix(build): use Ninja generator so cmake-rs stops clobbering MSVC /O2 
The cmake crate strips -O*/O* from CFLAGS and, with no explicit generator
on MSVC, overwrites CMAKE_C_FLAGS_RELEASE — so the previous CFLAGS=/O2
attempt could never work (verified: /DNDEBUG survived, /O2 stripped).
With CMAKE_GENERATOR=Ninja the override branch is skipped and CMake's
default Release flags (/O2 /Ob2 /DNDEBUG) reach ggml.
 2026-07-16 08:20:07 -05:00
iamdoubz 4e6c5a357e docs(bench): qualitative real-recording pass — P1 cleanest + best punctuation, W4 worst 2026-07-16 08:18:00 -05:00
iamdoubz 181af150e5 test(bench): qualitative meeting-audio pass — W2/W4/P1 on bank_deposits.wav 
62 s real recording (48 kHz stereo, downmixed/resampled by the production
reader). Reference .txt is a placeholder; WER column is meaningless here,
only the hypothesis text counts.
 2026-07-16 08:16:53 -05:00
iamdoubz d89b99db6c fix(build): restore /O2 /DNDEBUG for native deps clobbered by cmake-rs on MSVC 
cmake-rs hardcodes opt_level(0) and replaces CMake's Release C flags, so
ggml/whisper.cpp compiled at /Od — root cause of the W1 CPU-pathology
(RTF ~20) in docs/benchmarks/2026-07-parakeet-vs-whisper.md. cc-rs appends
env CFLAGS/CXXFLAGS last, restoring optimization for all native deps.
 2026-07-16 08:14:08 -05:00
iamdoubz 32e0d3c8d9 test(bench): raw benchmark CSVs (30-utterance LibriSpeech matrix + W1 non-vulkan rerun) 2026-07-16 00:42:26 -05:00
iamdoubz 21691d3756 docs(bench): parakeet-vs-whisper results — pre-registered criteria evaluated, per-rung recommendation 2026-07-16 00:42:25 -05:00
iamdoubz ca0563039d test(bench): p2-parakeet-dml contender (Parakeet on the iGPU via DirectML EP) 2026-07-15 23:14:47 -05:00
iamdoubz a682ffccdd test(bench): bench-directml feature for DirectML sherpa binaries (P2, never default) 2026-07-15 23:14:42 -05:00
iamdoubz 68632db6ef fix(bench): named Set-Content params — positional+NoNewline silently wrote nothing on PS 5.1 2026-07-15 21:45:46 -05:00
iamdoubz 297c412488 test(bench): sequential matrix runner (6 corpus + 3 live runs) 2026-07-15 21:42:28 -05:00
iamdoubz ad4495f220 test(bench): asr_bench harness — W1-W4 + Parakeet contenders, WER/RTF/CPU-seconds, live-window mode 2026-07-15 21:40:44 -05:00
iamdoubz 4dddcdfe1a test(bench): typeperf CPU/GPU/NPU counter capture script 2026-07-15 21:35:43 -05:00