mARTin-B78/tts-dgx-spark-faster-qwen3-tts
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Commit Graph

  • 6198f15f64 v6.8: Untrack generated VoiceDesign registry, add presets seed main mARTin-B78 2026-07-29 16:06:05 +0200
  • 9836261d28 Fix stale speaker embeddings and unregistered Voice Design voices mARTin-B78 2026-07-29 15:51:54 +0200
  • dbcbddedb4 Sync remote URLs: make GitHub primary, Gitea backup mARTin-B78 2026-06-26 11:24:32 +0200
  • 529599d98c Fix patch format mARTin-B78 2026-06-26 11:11:30 +0200
  • 3c389b0292 v6.7: Native speed control and zip word-level timestamps mARTin-B78 2026-06-26 11:11:13 +0200
  • 2ade7e1ca6 feat: add seed samples API endpoints mARTin-B78 2026-06-22 11:39:01 +0200
  • 79d8d2cd96 fix(customvoice): override max_seq_len to 1024 to prevent CUDA OOM mARTin-B78 2026-06-22 11:17:39 +0200
  • 377b9f16e5 feat: add /voice-seed API endpoint and update find_best_seed.py (v6.7.1) mARTin-B78 2026-06-21 14:53:31 +0200
  • 70fd8140e0 feat: add find_best_seed.py script for voice seed comparison mARTin-B78 2026-06-21 14:36:06 +0200
  • 602cf9e5df feat: deterministic per-voice RNG seed for consistent voice across requests (v6.7) mARTin-B78 2026-06-21 14:32:58 +0200
  • e58a6843b8 feat: eager background precomputation of speaker embeddings (v6.6) mARTin-B78 2026-06-21 09:35:42 +0200
  • 15f7fd63aa docs: clarify deterministic pt generation (v6.5.1) mARTin-B78 2026-06-21 00:53:05 +0200
  • 7a8d8d17ad feat: true zero-downtime voice hot-reloading (v6.5) mARTin-B78 2026-06-20 21:19:19 +0200
  • 4d7b2c5e0c feat: fully automate speaker embedding generation (v6.4) mARTin-B78 2026-06-20 20:52:53 +0200
  • 37274e52ee docs: add EN and DE reference voice samples mARTin-B78 2026-06-20 20:43:05 +0200
  • f482d07ad0 chore: ignore personal voices in config/speakers mARTin-B78 2026-06-20 13:50:14 +0200
  • 8f09b7e797 feat: precompute and load speaker embeddings (.pt) from voices.json mARTin-B78 2026-06-20 13:04:55 +0200
  • 789398c5df feat: introduce -streaming tag convention for Docker Hub mARTin-B78 2026-05-31 12:58:12 +0200
  • 4a199a883a docs: add v6.2 changelog entries for streaming fix and voice field persistence mARTin-B78 2026-05-30 21:05:43 +0200
  • 08be676685 fix: streaming service uses :latest image (fixes voice drift on port 8023) mARTin-B78 2026-05-30 21:04:00 +0200
  • 4ef2f893a7 fix: preserve user-added voice fields across container restarts mARTin-B78 2026-05-30 20:07:02 +0200
  • 8cdd81c5e5 docs: restore --max-seq-len in voiceclone compose, update v6 changelog mARTin-B78 2026-05-30 15:06:30 +0200
  • 40786314ef fix: regenerate patch against current upstream (instruct param removed) mARTin-B78 2026-05-30 14:43:43 +0200
  • 1fc2f9c909 fix: add --max-seq-len to patch and remove || true from Dockerfile mARTin-B78 2026-05-30 14:40:23 +0200
  • 3d8d444e0b fix: remove --max-seq-len from voiceclone command, fix Dockerfile comment mARTin-B78 2026-05-30 14:36:59 +0200
  • 9708eede71 refactor: consolidate Docker files into docker/ and merge streaming repo mARTin-B78 2026-05-30 14:31:25 +0200
  • 98405f681d fix: prevent voice drift and gender changes on long paragraphs (VoiceClone) mARTin-B78 2026-05-30 12:07:43 +0200
  • d04fa88853 refactor: load models asynchronously and add CUDA warmup for VoiceDesign and CustomVoice servers mARTin-B78 2026-05-26 23:07:51 +0200
  • eef48d11f5 docs: replace test beep with real William and Natasha voice samples mARTin-B78 2026-05-26 12:55:06 +0200
  • 9414d24b25 docs: add nvidia-ctk command per user feedback mARTin-B78 2026-05-26 12:25:14 +0200
  • d360164ad1 docs: add EN_M_Test dummy speaker and voices dir per user feedback mARTin-B78 2026-05-26 12:24:52 +0200
  • 41e6e734e1 docs: document streaming TTS backend mARTin-B78 2026-05-25 17:12:06 +0200
  • dfc2ca0e12 Point stack images to Docker Hub registry mARTin-B78 2026-05-25 09:50:53 +0200
  • c24bc263ec Clarify Qwen3 TTS service names mARTin-B78 2026-05-19 02:08:16 +0200
  • 2d3fdbfccc Align Qwen3 TTS stack with active voices mARTin-B78 2026-05-19 01:55:46 +0200
  • 486dd7c385 Add Qwen3 streaming TTS service mARTin-B78 2026-05-19 01:51:17 +0200
  • fc63f2c91d Ignore speakers backup, symlink, and workspace file mARTin-B78 2026-05-18 19:23:36 +0200
  • 8363dc7eb1 Add CustomVoice server, benchmark tool, and VoiceDesign API improvements mARTin-B78 2026-05-18 18:18:26 +0200
  • 295de3059d Add multi-source voice pipeline with VoiceDesign support mARTin-B78 2026-05-12 01:47:19 +0200
  • b7b917db66
         Update README.md mARTin 2026-05-25 13:21:47 +0200
  • 92e2381d0b Reduce latency: CUDA warmup, chunk_size=4, max-seq-len 2048 mARTin-B78 2026-05-11 20:43:21 +0200
  • 3f723b4032
         Comment out voice ID creation in generate_voices.py mARTin 2026-04-14 21:13:23 +0200
  • 60db1b181f Initial release: Faster-Qwen3-TTS for DGX Spark GB10 mARTin 2026-04-14 00:06:05 +0200