nodes/havvk-ComfyUI_AIIA
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • b8366be6ea fix(assets): replace seed_male_hq.wav with NotebookLM male voice (8s, 22050Hz) Hawk Lee 2026-02-15 15:45:11 +08:00
  • 28877f4e30 fix(stitcher): add 5ms fade-in/fade-out to eliminate pops at splice points Hawk Lee 2026-02-15 14:30:30 +08:00
  • f8ba18ef0c fix: auto-wrap each line with [1]: for multi-line plain text input Hawk Lee 2026-02-15 14:14:35 +08:00
  • feb7938ade revert(splitter): output pure text, VibeVoice auto-wrap handles [1]: prefix Hawk Lee 2026-02-15 14:12:02 +08:00
  • 3df8734b14 fix(splitter): add [1]: speaker prefix to output text for VibeVoice format Hawk Lee 2026-02-15 13:45:30 +08:00
  • d25f0e3ac9 fix(dialogue): use [N]: format and add voice_preset param for VibeVoice Hawk Lee 2026-02-15 13:26:33 +08:00
  • c0bf61185e fix: use official VibeVoice [N]: speaker format instead of Speaker N: Hawk Lee 2026-02-15 13:18:18 +08:00
  • ee51d51e38 fix(vibevoice): auto-wrap plain text as 'Speaker 1:' for processor compatibility Hawk Lee 2026-02-15 13:10:13 +08:00
  • 88710ae215 feat(vibevoice): add voice_preset dropdown for built-in reference voices Hawk Lee 2026-02-15 13:06:09 +08:00
  • c6c747b424 fix: restore overwritten remote files and add new ASR/Splitter/Stitcher nodes Hawk Lee 2026-02-15 12:20:55 +08:00
  • eff537d9dd feat: add ASR, Podcast Splitter, and Podcast Stitcher nodes for anti-leakage pipeline Hawk Lee 2026-02-15 12:14:45 +08:00
  • 37cd9096f2 Delete __pycache__ directory Hawk 2026-02-15 01:01:48 +08:00
  • 5343666e21 bump: version 1.11.2 -> 1.11.3 Hawk Lee 2026-02-04 20:05:23 +08:00
  • aa601cdeb0 fix(qwen): add fade-in/out to dialogue segments to prevent audio clicks Hawk Lee 2026-02-04 19:48:12 +08:00
  • 0859752cc9 feat(qwen): implement RMS normalization and reference audio volume scaling for Qwen3-TTS nodes Hawk Lee 2026-02-04 19:30:12 +08:00
  • 09d68fffa6 feat: Add more descriptive micro-expressions like 'With a hint of a smile' Hawk Lee 2026-02-04 19:07:11 +08:00
  • 9ae80b0be6 fix: Restore missing 'Gentle (温柔)' emotion to lists Hawk Lee 2026-02-04 19:04:19 +08:00
  • 1cbeb034de feat: Greatly expand Emotion and Expression lists with rich nuances and micro-expressions Hawk Lee 2026-02-04 19:01:43 +08:00
  • cf0907b944 feat: Add multilingual support (Chinese labels) to Emotion, Expression, and Dialect lists Hawk Lee 2026-02-04 18:57:29 +08:00
  • 79c4ec574b chore: Bump version to 1.11.2 Hawk Lee 2026-02-04 18:27:27 +08:00
  • b1b2f232f1 feat: Simplify AIIA Dialogue TTS node to use single Qwen model input (Bundle supported) Hawk Lee 2026-02-04 18:24:02 +08:00
  • b8686cedd2 fix: Fix model existence check in AIIA Qwen Dialogue TTS Hawk Lee 2026-02-04 17:14:28 +08:00
  • e30d664688 chore: Bump version to 1.11.1 and add Qwen3-TTS to description Hawk Lee 2026-02-04 17:09:00 +08:00
  • 567e67ae17 chore: Add debug logs for Qwen Dialogue TTS hang issue Hawk Lee 2026-02-04 17:03:01 +08:00
  • b3e7cbf91b fix: Make qwen_model input optional in AIIA Qwen Dialogue TTS Hawk Lee 2026-02-04 16:56:32 +08:00
  • 08374c4634 feat: Add segments_info output to AIIA Qwen Dialogue TTS Hawk Lee 2026-02-04 16:49:03 +08:00
  • eb22ed90ce diagnostics: temporarily disable custom labels JS Hawk Lee 2026-02-04 16:10:21 +08:00
  • 6c7e3cd469 fix: total ASCII stabilization and metadata index alignment Hawk Lee 2026-02-04 16:09:21 +08:00
  • 4dd27695fe fix: atomic UI reset to historical baseline (multiline, index alignment) Hawk Lee 2026-02-04 16:06:25 +08:00
  • cc057baf7f fix: total identity restoration and socket index alignment Hawk Lee 2026-02-04 16:04:28 +08:00
  • be4ae0e0ec fix: final surgical metadata alignment and NameError correction Hawk Lee 2026-02-04 16:01:11 +08:00
  • 54403ebd52 fix: total UI stabilization (stripped emojis, unified metadata) Hawk Lee 2026-02-04 15:56:16 +08:00
  • b06f5b16bc fix: emergency UI stability repair (stripped labels, fixed signatures, resolved NameErrors) Hawk Lee 2026-02-04 15:55:09 +08:00
  • 1cfdb3db73 fix: final precision structural restoration for Specialist node Hawk Lee 2026-02-04 15:45:58 +08:00
  • a2774db391 fix: remove duplicate optional block in Synthesis node Hawk Lee 2026-02-04 15:42:03 +08:00
  • a9126ab02f fix: total structural recovery to fix canvas socket shifts Hawk Lee 2026-02-04 15:41:01 +08:00
  • 5345a2324d fix: total structural restoration to fix canvas layout Hawk Lee 2026-02-04 15:38:29 +08:00
  • 0d65fea080 fix: total input order restoration to fix canvas sockets Hawk Lee 2026-02-04 15:36:23 +08:00
  • 1f9673781b fix: complete metadata alignment for Qwen3-TTS nodes Hawk Lee 2026-02-04 15:34:59 +08:00
  • 9d6816ebc4 fix: restore node inputs to fix canvas corruption Hawk Lee 2026-02-04 15:34:01 +08:00
  • 5b43ddd45f refactor: simplify Qwen3-TTS UI with Model Router and bundle support Hawk Lee 2026-02-04 15:22:22 +08:00
  • c4aefcf541 feat: add UI and console warnings for Qwen3-TTS Base model instruction limitations Hawk Lee 2026-02-04 14:52:16 +08:00
  • fc3720c723 docs: add Qwen3-TTS model-capability mapping table Hawk Lee 2026-02-04 14:30:30 +08:00
  • 94342c97ed docs: add dialect compatibility tips and update README Hawk Lee 2026-02-04 14:27:54 +08:00
  • 3d2d014e23 docs: add 0.6B download info and ModelScope support for Qwen3-TTS Hawk Lee 2026-02-04 14:05:19 +08:00
  • 178883eb96 feat: add dialect selection support for Qwen3-TTS and dialogue nodes Hawk Lee 2026-02-04 14:02:11 +08:00
  • 4dfa7a3432 fix: resolve batch_mode signature mismatch and improve max_batch_char robustness Hawk Lee 2026-02-04 11:46:16 +08:00
  • 08bf36ab4f docs: synchronize new parameters and emotion safeguards to README Hawk Lee 2026-02-04 11:00:19 +08:00
  • eb8527e934 fix: disable emotion tags for VibeVoice to prevent them being read aloud Hawk Lee 2026-02-04 10:58:21 +08:00
  • 8e0ca9abda feat: implement emotion compatibility check for non-expressive CosyVoice models Hawk Lee 2026-02-04 10:57:07 +08:00
  • 36514bd27d feat: add verbose console logging for text and instructions in dialogue nodes Hawk Lee 2026-02-04 10:54:46 +08:00
  • cda70248c0 feat: increase max_batch_char to 32768 for Qwen3-TTS long-context Hawk Lee 2026-02-04 10:53:14 +08:00
  • 0e49a1d850 feat: added specialized micro-expression controls for Qwen3-TTS Specialist node Hawk Lee 2026-02-04 10:49:40 +08:00
  • 481fa301c5 feat: expose max_batch_char to general Dialogue node and unify Qwen batching Hawk Lee 2026-02-04 10:42:39 +08:00
  • 4381966fe8 feat: implement batch maximization for Qwen3-TTS Dialogue node Hawk Lee 2026-02-04 10:40:45 +08:00
  • 6a8afb9b0d feat: unified emotion dropdowns for Qwen and general Dialogue nodes Hawk Lee 2026-02-04 10:04:05 +08:00
  • 3340d63766 feat: add emotion and expression dropdowns for Qwen3-TTS nodes Hawk Lee 2026-02-04 10:02:10 +08:00
  • 9288adecdd docs: add Qwen3-TTS instruction (emotion/style) control guide Hawk Lee 2026-02-04 09:59:33 +08:00
  • 51bd60c41f ui: rename x_vector_only to zero_shot_mode and enforce auto-fallback Hawk Lee 2026-02-04 09:56:26 +08:00
  • d27d6b6e5e feat: add reference text support for each speaker in Qwen Dialogue node Hawk Lee 2026-02-04 09:53:13 +08:00
  • 7f454d95f9 fix: auto-switch to zero-shot when ref_text is missing in VoiceClone mode Hawk Lee 2026-02-04 09:51:33 +08:00
  • 5d1f3fa371 fix: restore missing FUNCTION attribute in Qwen Specialist Dialogue node Hawk Lee 2026-02-04 09:49:40 +08:00
  • 76944fd165 fix: resolve qwen models from local folder_paths and log discovery Hawk Lee 2026-02-04 09:48:29 +08:00
  • 64217b1518 fix: resolve transformers 5.0.0 conflict and qwen-tts decorator bug Hawk Lee 2026-02-04 09:42:52 +08:00
  • b2acebdc4a feat: support high-quality fallback seeds in Qwen Specialist Dialogue node Hawk Lee 2026-02-04 09:29:56 +08:00
  • a76e46854e ui: implement automatic multi-line wrapping for static labels Hawk Lee 2026-02-04 09:26:34 +08:00
  • 8a60897c63 ui: implement elegant static labels for Qwen-TTS speaker guide using is_label Hawk Lee 2026-02-04 09:25:03 +08:00
  • 34fcb61a11 feat: support cfg_scale, temperature, top_k, top_p for Qwen3-TTS in all nodes Hawk Lee 2026-02-04 09:22:33 +08:00
  • 25f1bc7a73 ui: improved Qwen-TTS speaker selection with official list and UI notes Hawk Lee 2026-02-04 09:20:00 +08:00
  • 032f48aaa3 docs: add official speaker list to code comments Hawk Lee 2026-02-04 09:14:12 +08:00
  • 0ed49dbb1b docs: correct Qwen3-TTS preset speaker names Hawk Lee 2026-02-04 09:13:52 +08:00
  • 52d6e63042 feat: implement specialized AIIA_Qwen_Dialogue_TTS node for deep Qwen3 integration Hawk Lee 2026-02-04 09:11:47 +08:00
  • 98f615c603 feat: support multi-model routing for Qwen3-TTS in podcast nodes & clarify terminology Hawk Lee 2026-02-04 09:07:33 +08:00
  • 1d43237813 docs: complete manual download guide for all Qwen3-TTS model variants Hawk Lee 2026-02-04 08:59:48 +08:00
  • c8949e3b10 docs: add manual download guide for Qwen3-TTS Hawk Lee 2026-02-04 08:57:28 +08:00
  • 167edf8d20 feat: integrate Qwen3-TTS support into podcast nodes (v1.11.0) Hawk Lee 2026-02-04 08:54:08 +08:00
  • 9d069fc0ea v1.10.20: Enhanced subtitle calibration with Turn Protection (0.5s threshold) Hawk Lee 2026-02-04 00:22:29 +08:00
  • f264f140b7 v1.10.19: Fix subtitle timing with sequential priority logic Hawk Lee 2026-02-03 23:17:32 +08:00
  • 633ed75661 v1.10.17: Implement Speaker ID Mapping for subtitle calibration Hawk Lee 2026-02-03 01:29:00 +08:00
  • 9224e6c9f6 v1.10.16: Implement greedy speaker turn collection for subtitles Hawk Lee 2026-02-02 22:44:28 +08:00
  • d84ea55914 v1.10.15: Fix subtitle timing truncation bug Hawk Lee 2026-02-02 20:24:43 +08:00
  • 72391958b1 v1.10.14: Fix AttributeError 'should_stop' and improve cross-version compatibility Hawk Lee 2026-02-02 14:25:07 +08:00
  • de9eefcd64 v1.10.13: Fix Ditto Sampler interruption issue Hawk Lee 2026-02-02 14:20:52 +08:00
  • b5c2c80e0c v1.10.12: Add Text Debug Splicer node Hawk Lee 2026-02-02 13:54:19 +08:00
  • 81d7b23776 fix(subtitle): disable default bold style to make markdown emphasis visible (v1.10.11) Hawk Lee 2026-02-01 20:16:26 +08:00
  • c1639699d2 fix(vibevoice): implement smart audio fallback for multi-role support (v1.10.10) Hawk Lee 2026-02-01 16:37:06 +08:00
  • 50a2ec3346 fix(vibevoice): auto-normalize custom roles to Speaker N format (v1.10.9) Hawk Lee 2026-02-01 16:26:25 +08:00
  • d15c4cdc89 feat(subtitle): support markdown bold syntax in ASS generation (v1.10.8) Hawk Lee 2026-02-01 15:52:38 +08:00
  • 5f955c3c26 fix(subtitle): fix timeline alignment bug by allowing shared VAD chunks (v1.10.7) Hawk Lee 2026-02-01 15:16:33 +08:00
  • fce6175a24 test(subtitle): add comprehensive calibration tests and fix start-of-video fallback bug (v1.10.6) Hawk Lee 2026-02-01 01:45:26 +08:00
  • 5da3c028b8 fix(subtitle): implement speaker-centric matching to prevent cross-speaker calibration regressions (v1.10.5) Hawk Lee 2026-01-29 14:59:48 +08:00
  • 7e99c00efa fix(subtitle): fix calibration issues with out-of-order chunks and reversed timestamps (v1.10.4) Hawk Lee 2026-01-29 14:44:40 +08:00
  • 11ae63f3de fix(audio): improve compatibility with third-party extensions (LazyAudioMap support) Hawk Lee 2026-01-29 14:28:37 +08:00
  • f4babbe5ca fix(diarization): loosen audio validation to support float sample rates and list wrappers Hawk Lee 2026-01-29 14:19:50 +08:00
  • e968fc688b feat: implement subtitle calibration and subtitle-to-segments conversion (v1.10.3) Hawk Lee 2026-01-29 13:47:52 +08:00
  • fe8a8dda7b v1.10.1: Bump version Hawk Lee 2026-01-28 09:19:18 +08:00
  • ef2c434d01 v1.10.0: Add Speech Pitch Offset control and fix video combine UI bug Hawk Lee 2026-01-27 19:07:50 +08:00
  • ff187c9c00 v1.9.710: Continuous Vitality (Distance-to-Boundary Envelope) Hawk Lee 2026-01-27 02:13:07 +08:00
  • 2fce5c8164 feat: increase ASS subtitle font size to 40 and refine style Hawk Lee 2026-01-27 00:17:55 +08:00
  • 26b6a1d368 v1.9.700: Active Trajectory Planner & Vitality Boost Hawk Lee 2026-01-26 21:05:46 +08:00