Kazeia
/
kazeia
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • 2e6c55a068 feat(debug): encart temps = génération seule (sans la durée de lecture) chore/degraissage Kazeia Team 2026-06-21 14:48:28 +0200
  • 078804e2de feat(debug): encart « dernier tour » + verrou half-duplex entrée Kazeia Team 2026-06-21 14:42:28 +0200
  • 76ffbcb8fd fix(tts): robustesse lecture CosyVoice + pré-chauffage du cache voix Kazeia Team 2026-06-21 14:42:14 +0200
  • 7fffd3d9df feat(icons): nouvelles icônes adaptatives orbe (patient + admin) Kazeia Team 2026-06-21 08:52:25 +0200
  • 4d73a77ebd feat(voice): exclusivité au record-time + phrase de consentement enrichie Kazeia Team 2026-06-21 08:52:13 +0200
  • 8bcf9f8ebe fix(voices): /voices reflète l'inventaire CosyVoice (.cvps), pas les vestiges Qwen3 Kazeia Team 2026-06-21 08:51:56 +0200
  • 613c7efc66 fix(tts): repli sur la voix par défaut si le .cvps configuré est absent Kazeia Team 2026-06-18 23:40:42 +0200
  • ca530f6d85 fix(pipeline): régressions UI/STT post-bascule CosyVoice (orbe, texte, 1ère phrase) Kazeia Team 2026-06-18 22:51:35 +0200
  • ecda7b5715 chore(stt): retrait du backend STT "lib" (non utilisé/non validé) Kazeia Team 2026-06-18 10:56:58 +0200
  • a1c785bdf8 chore(cleanup): purge des traces legacy résiduelles (audit post-bascule) Kazeia Team 2026-06-18 10:48:03 +0200
  • a2fc01987b release(app): 0.2.3 / versionCode 15 (APK lean sans natifs JNA) Kazeia Team 2026-06-18 09:38:51 +0200
  • 22dbd2145f build: exclure les natifs JNA cross-plateforme (-2,3 Mo, APK = binaires utiles) Kazeia Team 2026-06-18 09:20:13 +0200
  • 66e81f0668 refactor(tts): bascule CosyVoice = seul TTS + purge legacy (≈322 Mo libs) Kazeia Team 2026-06-18 09:07:41 +0200
  • 3d002318d5 feat(updates): MAJ manuelles pilotées depuis l'admin (plus d'auto-update) Kazeia Team 2026-06-18 08:27:41 +0200
  • a2f82f69f9 release(app): refonte 0.2.0 (versionCode 12) Kazeia Team 2026-06-18 08:09:06 +0200
  • 577aad3b2f fix(dist): hasCoreModels() reflète le cœur post-refonte (GGUF speaker + Whisper) Kazeia Team 2026-06-18 07:52:22 +0200
  • 471fb5d5fb feat(admin): presets d'échantillonnage LLM (Speaker/Thinker) — app-side complet Kazeia Team 2026-06-18 05:59:18 +0200
  • b11d8a96d4 diag(bench): option .pte multi-contexte en premier (bench_llm_pte_first) Kazeia Team 2026-06-18 00:13:54 +0200
  • 93a3953614 feat(admin): toggle "mode debug" — masque le bouton logs/métriques côté patient Kazeia Team 2026-06-17 23:43:50 +0200
  • d1baacd096 refonte(llm): Phase 3 — archivage code mort (Genie + EngineLlmAdapter) Kazeia Team 2026-06-17 23:28:14 +0200
  • 0108f47c09 refonte(llm): NPU .pte in-app sans root (ADSP_LIBRARY_PATH) + normalisation think-block Kazeia Team 2026-06-17 23:00:31 +0200
  • 0a72654093 refonte(llm): moteur LLM unifié GGUF + .pte NPU (UnifiedLlmAdapter) Kazeia Team 2026-06-17 22:47:32 +0200
  • 3b58fcb467 feat(tts): CosyVoice rapide (hiftf16, RTF 0.88 + cache) + guard init — v0.1.10 main Kazeia Team 2026-06-15 23:50:42 +0200
  • db0e56b26c perf(tts): NE PAS streamer LLM→CosyVoice (contention CPU) — documenté — v0.1.9 Kazeia Team 2026-06-15 12:29:25 +0200
  • b822974b99 fix(llm): sessionReset natif corrigé (archi hybride) — reset de borne propre — v0.1.8 Kazeia Team 2026-06-15 11:26:00 +0200
  • 8d548b1ca6 feat(llm): session cache-préfixe KV + streaming + mémoire conv. — v0.1.7 Kazeia Team 2026-06-15 11:12:11 +0200
  • 2acf3922c9 feat(llm): défaut kazeia-engine GGUF (lib), plus de .pte au runtime — v0.1.6 Kazeia Team 2026-06-15 08:15:27 +0200
  • 359252bf5e feat(voice): voix pilotée par profil (app admin) — retrait du sélecteur patient — v0.1.5 Kazeia Team 2026-06-15 07:51:42 +0200
  • cca38ef50f feat(tts): intégrer libs CosyVoice3 auto-contenues + validation device — v0.1.4 Kazeia Team 2026-06-14 21:31:42 +0200
  • ed61737199 feat(tts): adapter CosyVoice3 + dispatch ttsEngine="cosyvoice" (câblage, libs en attente) Kazeia Team 2026-06-14 17:53:30 +0200
  • 16d56bfd4d fix: défaut ttsEngine="lib" (crash-loop install fraîche) + ragCorpusDir externe-d'abord — v0.1.3 Kazeia Team 2026-06-11 23:32:16 +0200
  • 747848a475 feat(rag): syncDir — synchro corpus fichiers→base à chaque démarrage Kazeia Team 2026-06-11 22:51:30 +0200
  • 3761e03d52 feat(rag): corpus psychoéducation FR étendu à 18 fiches (corpus-seed-r2) Kazeia Team 2026-06-11 22:49:36 +0200
  • d7e90c16c4 fix(safety): rappel « mieux mort » + regex normalize compilées une fois Kazeia Team 2026-06-11 16:22:27 +0200
  • 7669a79f4f fix(rag): ingestion atomique + build/teardown sérialisés (code-review) Kazeia Team 2026-06-11 16:10:29 +0200
  • e1d9c94751 fix(safety,rag): correctifs critiques code-review Kazeia Team 2026-06-11 16:07:51 +0200
  • 32669ae039 docs: emplacement Nextcloud du tarball jniLibs + recette clone→build complète Kazeia Team 2026-06-11 15:45:03 +0200
  • d47206bff8 release: bump v0.1.1 (versionCode 2) — premier cycle OTA réel validé Kazeia Team 2026-06-11 11:20:39 +0200
  • 9c6f4e8838 build: reproductibilité des jniLibs — manifest sha256 + script + garde release Kazeia Team 2026-06-11 09:31:00 +0200
  • e5b48e4e0f test: tests unitaires JVM — CrisisGuard (sécurité vitale) + VectorIndex + Chunker Kazeia Team 2026-06-11 08:44:09 +0200
  • d834ea4955 docs: RELEASE_SIGNING à jour — keystore versionné dans le dépôt privé comme sauvegarde Kazeia Team 2026-06-11 08:38:16 +0200
  • d936b0d55d build: versionner le keystore release dans le dépôt privé (sauvegarde) Kazeia Team 2026-06-11 08:37:34 +0200
  • 93f09c87d3 build: signature release (keystore partagé patient+admin) + doc de gestion Kazeia Team 2026-06-11 08:32:29 +0200
  • fd9e04a0cb feat(rag): gestionnaire RAG complet dans l'app admin Kazeia Team 2026-06-11 07:53:31 +0200
  • 4508b8fe2e feat(safety): garde-fous cliniques + filet de crise déterministe Kazeia Team 2026-06-10 23:32:38 +0200
  • acaf0cf2cf feat(rag): écran admin d'ingestion du corpus (provider /rag + réindexation live) Kazeia Team 2026-06-10 13:57:32 +0200
  • 62a9803fdd feat(rag): intégration RAG dans le flux patient (corpus + ingestion + injection live) Kazeia Team 2026-06-10 12:12:21 +0200
  • 95a7a25335 feat(rag): valider EngineEmbedder e5 in-app (E2E réel) + préfixe passage Kazeia Team 2026-06-10 12:02:44 +0200
  • 3eab29beb2 feat(rag): module RAG on-device (indépendant de l'engine), flag OFF par défaut Kazeia Team 2026-06-09 22:09:57 +0200
  • 6be3e47148 docs: spec embeddings RAG pour kazeia-engine + whitelist docs/ Kazeia Team 2026-06-09 21:54:32 +0200
  • 4498e5ac7a chore: dégraissage du dépôt — suivi restreint à l'app mobile (liste blanche) Kazeia Team 2026-06-09 21:21:00 +0200
  • 6f45a75197 chore: snapshot migration moteur lib (préservation avant dégraissage) Kazeia Team 2026-06-09 21:19:24 +0200
  • 06b333f0f0 feat(speaker): rendre les Speakers GGUF sélectionnables in-app (Qwen3.5-4B, Qwen2.5-7B) feat/speaker-gguf-selectable Kazeia Team 2026-06-07 23:14:06 +0200
  • 8a3151681d llm: add llama.cpp backend (LlamaCppLlmEngine) for Qwen3.5-4B Kazeia Team 2026-04-23 08:07:07 +0200
  • db281002d9 scripts: export per-voice prefix/suffix embeddings Kazeia Team 2026-04-15 00:09:23 +0200
  • c2f7859dfe UI+TTS: voice hot-swap + typing dots + emoji stripping Kazeia Team 2026-04-14 23:55:07 +0200
  • b5b13780f7 UI: whole-sphere Fourier-mode deformation during speech Kazeia Team 2026-04-14 23:47:30 +0200
  • 2fe46e0f15 Fix seg-2 audio dropout + switch spectrum from bars to Bézier lines Kazeia Team 2026-04-14 23:42:43 +0200
  • 06dcd76dcb UI: large central orb w/ spectrum-inside + per-voice palette Kazeia Team 2026-04-14 23:33:38 +0200
  • 8939c680b2 UI: épuré audio-reactive orb visualizer — replaces 3D avatar for MVP Kazeia Team 2026-04-14 23:20:15 +0200
  • f17131aefb UI: reveal Kazeia reply in sync with TTS audio (per-sentence, per-word) Kazeia Team 2026-04-14 22:58:18 +0200
  • 6a958c1a10 Revert MemoryOptimizer — reclaim wasn't worth the footprint Kazeia Team 2026-04-14 22:46:58 +0200
  • 751e3e0868 memory: periodic sweep + expand kill list (photos, calendar, contacts, vending, tachyon…) Kazeia Team 2026-04-14 22:44:09 +0200
  • 39babcb158 TTS+audio+memory: ColorOS playback fixes + kill-background reclaim Kazeia Team 2026-04-14 22:37:20 +0200
  • 0632db1ee0 UI: drop Magisk prompt — ResourceMonitor stops probing su Kazeia Team 2026-04-14 18:35:18 +0200
  • 10fd10fd90 TTS: overlap CP↔BigVGAN — first audio 14.5s → 10.9s per segment Kazeia Team 2026-04-14 16:22:15 +0200
  • 67de8d4767 LLM: enable hybrid-mode export via num_sharding=1 — TTFT 2.9s → 113ms Kazeia Team 2026-04-14 15:08:31 +0200
  • a41619ed67 TTS: keep BigVGAN on CPU after GPU regression; LLM filter strips more tags Kazeia Team 2026-04-14 13:48:37 +0200
  • f4b15a72a7 LLM JNI: auto-detect eval_mode from .pte methods (kv-only vs hybrid) Kazeia Team 2026-04-14 12:45:10 +0200
  • 3d435f9cdd LLM: trim system prompt to drop ~27 prefill tokens (-1.3s TTFT) Kazeia Team 2026-04-14 12:16:11 +0200
  • 7dc6704e95 docs: add before/after performance comparison to no-root report Kazeia Team 2026-04-14 11:37:15 +0200
  • 6c7746c5d0 docs: add post-mortem to no-root report — issue resolved Kazeia Team 2026-04-14 11:19:27 +0200
  • b57719fa5e LLM: filter <think> tokens out of the streaming TTS path Kazeia Team 2026-04-14 11:16:08 +0200
  • f32b5ddfdd LLM no-root: validate end-to-end pipeline, fix kv_io_bit_width detection Kazeia Team 2026-04-14 11:11:23 +0200
  • 809a6d4fed LLM no-root: migrate to in-process LlmModule (JNI) — zero su calls Kazeia Team 2026-04-14 10:39:50 +0200
  • 6e6a2d9f82 Baseline before no-root migration: working state with root LLM backup/pre-no-root-migration Kazeia Team 2026-04-14 08:19:36 +0200
  • 364016b7b8 LLM+TTS: short-response system prompt, PTE streaming fallback Kazeia Team 2026-04-14 00:17:08 +0200
  • 9930bfa392 LLM: enable Qwen3-4B NPU (21 tok/s) in service pipeline Kazeia Team 2026-04-13 23:00:25 +0200
  • 19f934af25 LLM NPU: Qwen3-4B QNN export patches + deployment notes Kazeia Team 2026-04-13 22:56:42 +0200
  • f548e02283 TTS: dynamic EOS-rank boost terminates generation cleanly across voices Kazeia Team 2026-04-13 14:13:04 +0200
  • c25040a780 TTS: conditional tail-trim + export script accepts voice path arg Kazeia Team 2026-04-13 11:32:33 +0200
  • 0833d1bd21 TTS: route all synthesizeAndPlay calls through Stage 3 streaming session Kazeia Team 2026-04-13 11:12:14 +0200
  • 2f07901ff3 TTS Stage 3: LLM stream → sentence split → TTS session → shared AudioTrack Kazeia Team 2026-04-13 10:52:46 +0200
  • 7f1a44c23d TTS Stage 2: on-device voice-cloning TTS for arbitrary text Kazeia Team 2026-04-13 10:12:09 +0200
  • 5e416713ce TTS Stage 1 streaming: play each segment the moment it's decoded Kazeia Team 2026-04-13 08:43:30 +0200
  • de878ddf5c TTS tremor investigation: identify cross-arch numerical floor, gate diag flags Kazeia Team 2026-04-13 00:15:14 +0200
  • ee186e9049 Auto-segmentation for long texts + dynamic pipeline Kazeia Team 2026-04-10 00:08:59 +0200
  • 199bc4fbc9 Full native C++ TTS validated on short + long phrases Kazeia Team 2026-04-09 23:51:05 +0200
  • dafbe2a52b FULL NATIVE C++ TTS pipeline — any text, perfect quality Kazeia Team 2026-04-09 23:39:06 +0200
  • 09d36f2025 Root cause found + on-device embed capture + KV=100 restored Kazeia Team 2026-04-09 23:00:37 +0200
  • 3dcf73aa38 Restore KV=100 + fix as-is embeds + multi-segment support Kazeia Team 2026-04-09 22:26:20 +0200
  • 10a3904d7d Multi-segment TTS for long text: split → generate → concatenate Kazeia Team 2026-04-09 14:34:05 +0200
  • 24157c0a68 Fix: use pre-computed embeds as-is (no double codec_sum) Kazeia Team 2026-04-09 14:10:23 +0200
  • f6df1738c5 Add prepare_tts_embeds.py for any text + codec_sum fix Kazeia Team 2026-04-09 14:05:42 +0200
  • 173606dae7 Stable: decoder 8T optimization + restore pre-computed embeds Kazeia Team 2026-04-09 13:42:02 +0200
  • 42bbb96fd8 Optimize decoder: BigVGAN 8T, small models 4T → RTF 1.26 Kazeia Team 2026-04-09 13:00:05 +0200
  • a688edc9ec Reduce talker KV_LEN 100→64: saves 148ms (RTF 1.31) Kazeia Team 2026-04-09 12:47:30 +0200
  • 4dcc4bb8b3 Fix KV buffer + revert HTP decoder (BigVGAN too complex for HTP) Kazeia Team 2026-04-09 12:37:50 +0200
  • 985fd9cff9 Direct output→input KV copy: RTF 1.51 → 1.31 Kazeia Team 2026-04-09 12:23:45 +0200
  • 14f7e5b05f Optimize CP+talker: eliminate prepare_input_tensors per step Kazeia Team 2026-04-09 12:16:38 +0200