Nighthawk/Qwen3-TTS-streaming
1
0
Fork 0
mirror of https://github.com/Nighthawk42/Qwen3-TTS-streaming.git synced 2026-08-30 10:02:26 +00:00
Code Issues Packages Projects Releases Wiki Activity
57 Commits 1 Branch 0 Tags
main
Commit Graph
17 Commits
Author SHA1 Message Date
Kedara Studios b8db6ff1ce update README.md: list experimental branch features; GPU-resident repetition penalty and batch compaction 2026-02-12 18:54:48 +01:00
Kedara Studios a757bf2bf0 update README.md -> move to experimental branch 2026-02-12 02:36:47 +01:00
Kedara Studios 81275c6895 Update README.md 2026-02-10 11:21:01 +01:00
Kedara Studios 7e616bbf9f docs: reorder README sections and update repetition_penalty description 2026-02-08 17:45:00 +01:00
Kedara Studios fa20ec2dc3 docs: add batch streaming section to README with usage example 
Document the new batch_stream_generate_voice_clone() method including
a usage example showing voice prompt broadcasting, per-item chunk
accumulation, and WAV output. Add entry to the "Added in this fork" list.
 2026-02-08 17:36:34 +01:00
rekuenkdr f88566ac7e Merge branch 'main' into fix/repetition-penalty-and-finetune-fixes 2026-02-07 10:42:24 +01:00
Kedara Studios a3f95c5e65 docs: add repetition_penalty param to README with streaming context 2026-02-07 10:36:34 +01:00
Kedara Studios 909e1089be feat(audio): make overlap_samples fully configurable 
Allow users to set any overlap_samples value including 0 to disable
crossfade blending entirely. Renamed MIN_BLEND_SAMPLES to
DEFAULT_BLEND_SAMPLES and removed the enforced minimum.
 2026-02-03 08:48:13 +01:00
Kedara Studios 96137074c8 fix(audio): Hann window crossfade and boundary click prevention 
- Replace linear crossfade with Hann window for smoother transitions
- Add MIN_BLEND_SAMPLES (512) as floor for blend region
- Add Hann fade-in on first chunk to prevent startup pop
- Add Hann fade-out on final chunk to prevent ending pop
- Trim chunk tails before emission to prevent echo artifacts
- Update README with audio quality fixes documentation
 2026-02-02 23:38:16 +01:00
Kedara Studios 7bf2c43a2e feat: add two-phase streaming for reduced first-chunk latency 
- Phase 1: aggressive emit/decode settings for fast first chunk
- Phase 2: stable settings with optimized decode for quality
- New parameters: first_chunk_emit_every, first_chunk_decode_window, first_chunk_frames
- Phase 1 disables torch.compile to allow flexible window sizes
 2026-02-02 20:27:15 +01:00
dffdeeq 01b51f0cb1 README update 2026-01-28 02:26:50 +04:00
Andrew 648b7100bd benchmarks update 
Fixed formatting issues and improved clarity in the README.
 2026-01-26 17:54:30 +04:00
Andrew e03d44f3f4 README.md update 2026-01-25 23:45:27 +04:00
Xiong Wang 1ab0dd7535 update: update readme 2026-01-26 00:11:02 +08:00
Xiong Wang 3b30a4e509 update: update readme 2026-01-24 21:44:35 +08:00
Xiong Wang 8a98526da0 fix and update: fix finetuning/prepare_data.py and update citation 2026-01-23 13:19:19 +08:00
Xiong Wang d8146d5305 Initial commit 2026-01-22 17:57:04 +08:00