Commit graph

Select branches

Hide pull requests

fix-models-docs-branch

main

Mono Color

• 585571c609
  speech.py --engines: default to F5 only; 'all' opts into every engine main russell@unturf.com 2026-06-10 12:39:28 -04:00
• a0d9e9a966
  add --engines CLI flag to allowlist TTS backends russell@unturf.com 2026-06-10 12:37:58 -04:00
• d4ff4e8da4 speech.py: disable tts-1-qwen registration — free VRAM for llama-qwen LLM timehexon 2026-06-07 08:39:49 -04:00
• 33bd85f1cb
  CLAUDE.md: document mispronunciation fix workflow russell@unturf.com 2026-06-05 19:34:24 -04:00
• 9f54ae9053
  pre_process_map: drop unused Providence stub + stale A/B comment russell@unturf.com 2026-06-05 19:29:14 -04:00
• 09ae91a26a
  pre_process_map: Provenance -> prahvanans (enable, per fox) russell@unturf.com 2026-06-05 19:28:37 -04:00
• d6cd1614ca
  pre_process_map: stage Provenance + Providence fixes, commented for A/B russell@unturf.com 2026-06-05 19:21:39 -04:00
• c161a95eef
  pre_process_map: target Provenance (not Providence) russell@unturf.com 2026-06-05 19:18:32 -04:00
• 5f2f4bfcc8
  pre_process_map: Providence -> Prav-uh-dence (fix vowel) russell@unturf.com 2026-06-05 18:40:58 -04:00
• 221756e501
  pre_process_map: track config file in repo; fix Providence pronunciation russell@unturf.com 2026-06-05 18:03:31 -04:00
• 468cd1dfc0 Makefile: fold sox into apt-deps (now a required dep) timehexon 2026-06-05 12:33:10 -04:00
• 5e41ab6779 Makefile: add apt-deps + apt-deps-sox targets for venv path timehexon 2026-06-05 12:04:24 -04:00
• 3b58f0c867
  F5-TTS: add SSE streaming mode with per-sentence audio + timing russell@unturf.com 2026-05-29 15:15:32 -04:00
• b503c4d5d6
  F5-TTS: add opt-in sentence timestamps mode russell@unturf.com 2026-05-29 12:29:18 -04:00
• e511d4d105 whisper_refs: fill foxhop ref_text from actual recording timehexon 2026-05-25 06:23:18 -04:00
• 9f15d01569
  Add voice 42: foxhop (self-recorded) russell@unturf.com 2026-05-25 06:21:31 -04:00
• c5247cd993
  whisper_refs: normalize transcripts (sentence-case, strip quote artifacts) russell@unturf.com 2026-05-24 16:02:18 -04:00
• 72a9553047
  CLAUDE.md: replace stale 3090-ai+docker prod docs with 4090+bare-python reality russell@unturf.com 2026-05-24 16:02:08 -04:00
• 4ac192585d F5-TTS: replace LibriSpeech ground-truth labels with real Whisper transcripts blackops 2026-05-24 14:47:25 -04:00
• e9c978b051
  endpoints: filter /v1/models and /v1/voices by engine availability flags russell@unturf.com 2026-05-24 11:58:35 -04:00
• f25731ca08
  Add make whisper-refs — re-transcribe cloned-voices/*.wav with whisper-large-v3 russell@unturf.com 2026-05-24 11:56:15 -04:00
• 3a76ba6748
  CLAUDE.md: refine our 'avoid the' style rule (match unsandbox shard) russell@unturf.com 2026-05-24 11:56:10 -04:00
• 831b937fd1
  F5-TTS: per-chunk silence trim + fade (port from VoiceClone) + restore per-sentence streaming russell@unturf.com 2026-05-24 11:39:49 -04:00
• 015bbdfad6 F5-TTS: Whisper-transcribed ref_texts with full punctuation for all 40 cloned voices timehexon 2026-05-24 11:34:12 -04:00
• 960fda4f5c
  F5-TTS: sentence-case + period the LibriSpeech ref_texts for cleaner ref-to-gen transition russell@unturf.com 2026-05-24 11:26:34 -04:00
• fc5af2b653
  F5-TTS: drop per-sentence split — single infer() call avoids ref-to-gen artifact per chunk russell@unturf.com 2026-05-24 11:12:57 -04:00
• 482b17960c
  F5-TTS engine consistency sweep (Makefile, startup, README, docs) russell@unturf.com 2026-05-24 08:24:38 -04:00
• 2df34f85cc
  Add F5-TTS as tts-1-f5 engine (additive, alongside tts-1-qwen) russell@unturf.com 2026-05-23 13:24:25 -04:00
• bfc5260dfc we update gitignore russell@unturf.com 2026-04-03 14:13:37 -04:00
• 074cbbd22f style: avoid "the", use "our" — writing style rule + sweep russell@unturf.com 2026-03-31 13:20:22 -04:00
• d4d6344326 Replace "AI" with "machine learning" in CLAUDE.md timehexon 2026-02-02 19:56:24 +00:00
• 7e77c3452d Add voices-qwen Makefile target to download cloned voice samples russell@unturf.com 2026-01-27 17:31:15 -05:00
• f58da54fa3 Expand to all 40 test-clean voices with idempotent registry fox 2026-01-27 15:30:42 -05:00
• 69d039d1a5 Add datasets and torchcodec to requirements.txt russell@unturf.com 2026-01-27 15:15:09 -05:00
• 3f3918fed2 Simplify CLAUDE.md: remove SSH/rsync references, local-only workflow russell@unturf.com 2026-01-27 13:59:52 -05:00
• c3ac2def48 Add idempotent voice registry system for permanent speaker-to-name assignments russell@unturf.com 2026-01-27 13:59:39 -05:00
• af5c81928d Regenerate voices with upstream-verified genders from SPEAKERS.TXT fox 2026-01-27 13:22:26 -05:00
• 35b083ed34 Fetch speaker genders from upstream LibriSpeech SPEAKERS.TXT russell@unturf.com 2026-01-27 13:17:52 -05:00
• b77af6d603 Fix voice gender assignments - verified against LibriSpeech SPEAKERS.TXT russell@unturf.com 2026-01-27 13:13:07 -05:00
• c56b8cb24e Simplify Makefile to run locally, remove all remote SSH/rsync russell@unturf.com 2026-01-27 12:51:08 -05:00
• ebfa0319d1 Replace voices with 21 distinct gendered speakers from LibriSpeech russell@unturf. 2026-01-27 12:41:18 -05:00
• cd60e31aa2 Use gendered voice names: 11 female + 10 male = 21 distinct voices russell@unturf.com 2026-01-27 12:07:31 -05:00
• 4c5b42c062 Mount cloned-voices directory in container russell@unturf.com 2026-01-27 11:56:17 -05:00
• 648281cdb6 Update diverse voices script with correct voice names russell@unturf.com 2026-01-27 11:53:33 -05:00
• aa8af56835 Fix remote paths in CLAUDE.md, add tmux-hosts discovery russell@unturf.com 2026-01-27 09:32:27 -05:00
• 16b281bab4 Add webm response format (opus in webm container) russell@unturf.com 2026-01-27 09:28:38 -05:00
• 8f7f1318a1 Split on every sentence for streaming (no combining) russell@unturf.com 2026-01-26 19:52:41 -05:00
• 0a0d023517 Fix sentence splitter to split on every sentence boundary russell@unturf.com 2026-01-26 19:41:44 -05:00
• 802eaf2b29 Add sentence-by-sentence streaming for Qwen TTS russell@unturf.com 2026-01-26 19:32:51 -05:00
• a148088cb0 Fix Qwen TTS deadlock, reduce workers to 1 for GPU russell@unturf.com 2026-01-26 19:14:19 -05:00
• cac40d75d3 Set default workers to 4 to prevent server lockup russell@unturf.com 2026-01-26 18:23:51 -05:00
• 5e5e7936f1 Add 20 cloned voice samples for Qwen3-TTS russell@unturf.com 2026-01-26 17:44:03 -05:00
• 02b4e7aaf7 Add 20 diverse voice samples for Qwen3-TTS russell@unturf.com 2026-01-26 16:50:54 -05:00
• 6b4f66dcf5 Use LJ Speech sample for voice cloning (Alibaba Cloud URL blocked) russell@unturf.com 2026-01-26 16:28:39 -05:00
• ad6a4d4990 Fix docker-compose.yml for older docker-compose versions russell@unturf.com 2026-01-26 14:25:06 -05:00
• 0159f1f216 Add configurable WORKERS env var, default to 1 for GPU models russell@unturf.com 2026-01-26 13:16:44 -05:00
• f299b43f1a Fix qwen-tts version constraint (0.0.5 is latest) russell@unturf.com 2026-01-26 11:14:26 -05:00
• 381ba3b462 Pre-download Qwen3-TTS model on container startup russell@unturf.com 2026-01-26 11:07:30 -05:00
• b315659be6 Make Qwen3-TTS the default engine, add CPU-only docker support russell@unturf.com 2026-01-26 10:40:29 -05:00
• 4a019cf897 Add detailed AGPL v3 license obligations documentation russell@unturf.com 2026-01-26 10:39:23 -05:00
• 99bc6bf014 renamed: docs/CLAUDE.md -> CLAUDE.md russell@unturf.com 2026-01-26 10:00:07 -05:00
• 058ad5840b Fix docker-compose image references to use local builds Russell Ballestrini 2025-12-13 10:53:36 -05:00
• 7559e56d0c Standardize project naming to uncloseai-speech across all files Russell Ballestrini 2025-11-10 05:23:34 -05:00
• da8e960b2d Fix Kokoro defaulting to CPU in worker processes Russell Ballestrini 2025-11-10 04:22:53 -05:00
• ae958d1bb6 Enable GPU acceleration for Kokoro TTS Russell Ballestrini 2025-11-10 04:14:46 -05:00
• 4576afac39 Fix UnboundLocalError in cleanup function Russell Ballestrini 2025-11-10 04:01:39 -05:00
• bb9823f6d0 🦝 Move XTTS imports to module level for worker processes Russell Ballestrini 2025-11-09 18:40:38 -05:00
• 3887e9b850 🦝 Streamline CLAUDE.md - reference guide not changelog Russell Ballestrini 2025-11-09 18:29:24 -05:00
• 187121558e 📚 Update docs with hydration results and multiprocess fixes Russell Ballestrini 2025-11-09 18:27:48 -05:00
• be374409d6 🦝 Fix args being None in worker processes Russell Ballestrini 2025-11-09 18:22:19 -05:00
• bc2071900c 🦝 Fix voice cache initialization in multiprocess workers Russell Ballestrini 2025-11-09 18:13:52 -05:00
• 549a35d613 🦝 Add working hydrate and load-test targets Russell Ballestrini 2025-11-09 17:33:58 -05:00
• aeebb69a8c 🦝 Fix uvicorn workers with import string Russell Ballestrini 2025-11-09 17:10:08 -05:00
• 459e8d5896 🦝 Fix concurrency with multiprocess workers + semaphores Russell Ballestrini 2025-11-09 16:44:06 -05:00
• 650ae49f65 🦝 Fix blocking model loads - enable concurrent TTS requests Russell Ballestrini 2025-11-09 16:05:12 -05:00
• 21e8f27519 🦝 Fix /v1/voices timeout by caching response data at startup Russell Ballestrini 2025-11-09 15:58:01 -05:00
• ac305c31d7 Remove abandoned GitHub mirror from README Russell Ballestrini 2025-11-09 15:24:22 -05:00
• 066b3e9f08 Merge remote changes (keep our auto-detection and alias cleanup) Russell Ballestrini 2025-11-09 15:23:11 -05:00
• 04e4e5ae84 Update repository links and prefer Makefile workflow Russell Ballestrini 2025-11-09 15:19:59 -05:00
• 32393c7665 Remove misleading Silero OpenAI voice aliases Russell Ballestrini 2025-11-09 15:08:06 -05:00
• 3df4ff02a7 Merge branch 'claude/implement-models-docs-011CUxXuNMytPjEr5vsvcboo' into 'main' Russell Ballestrini 2025-11-09 19:48:45 +00:00
• 1b4d231556 Add comprehensive TTS model documentation and research Russell Ballestrini 2025-11-09 19:48:45 +00:00
• d8b9a06b45 Add voice-based model auto-detection and voice discovery endpoint Russell Ballestrini 2025-11-09 14:45:06 -05:00
• 9b5caadb8f Fix Kokoro TTS integration - correct KPipeline API Russell Ballestrini 2025-11-09 14:19:36 -05:00
• f8d46e92d5 Update documentation for Silero and Kokoro integrations Russell Ballestrini 2025-11-09 13:55:30 -05:00
• 372c6a5d3f Add /v1/models endpoint for voice discovery Russell Ballestrini 2025-11-09 13:51:11 -05:00
• 603a211f47 Add /v1/models endpoint for voice discovery Russell Ballestrini 2025-11-09 13:39:32 -05:00
• d48fa6b29c Integrate Kokoro TTS as tts-1-kokoro model Russell Ballestrini 2025-11-09 13:37:48 -05:00
• 1a27597d94
  Add comprehensive testing guide for Silero TTS deployment Claude 2025-11-09 18:34:40 +00:00
• 20241632ea Fix Silero multilingual support with proper model loading Russell Ballestrini 2025-11-09 13:29:52 -05:00
• a8865564ae Map all available Piper voices and expand Makefile downloads Russell Ballestrini 2025-11-09 13:16:22 -05:00
• 01e51b08b5 🦝 Raccoon Mission: Silero TTS integration complete with 140 voices Russell Ballestrini 2025-11-09 12:39:12 -05:00
• 4deedb9539 Fix syntax error in speech.py and document Chatterbox dependency conflict Russell Ballestrini 2025-11-09 10:49:34 -05:00
• 848c2c6cb5 Integrate Silero TTS and add infrastructure for Chatterbox/Kokoro Claude 2025-11-09 15:31:47 +00:00
• 2d1e1b344f Add comprehensive TTS model documentation and research Claude 2025-11-09 15:21:54 +00:00
• 6e2b0cbf32 Document Chatterbox dependency conflict in requirements.txt fix-models-docs-branch Russell Ballestrini 2025-11-09 10:43:15 -05:00
• cd4248045e Update MODELS.md with comprehensive TTS engine documentation Russell Ballestrini 2025-11-09 10:37:57 -05:00
• 759c5440b7 Update README with Raccoon Mission details and mirrors Russell Ballestrini 2025-11-09 09:56:31 -05:00
• 7b11e4c336 Add GitHub remote and push-all Makefile target Russell Ballestrini 2025-11-09 09:55:42 -05:00
• d6c9654816 Copy scripts/ directory into Docker container Russell Ballestrini 2025-11-09 09:52:57 -05:00