-
Notifications
You must be signed in to change notification settings - Fork 1.6k
Pull requests: QwenLM/Qwen3-TTS
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Fix speaker_encoder silent decay under AdamW weight decay in SFT
#351
opened Jul 24, 2026 by
MR-AI-HuiMin
Loading…
4 tasks
Fix 0.6B fine-tuning crash: project text embedding before adding to c…
#336
opened Jun 30, 2026 by
kru2710shna
Loading…
fix: Set dtype on nested model config to avoid flash attention warning
#309
opened Apr 27, 2026 by
mitchsayre
Loading…
fix: provide TensorBoard logging dir in finetuning script
#295
opened Apr 14, 2026 by
ElhamDevelopmentStudio
Loading…
Pre-wire instruct support in generate_voice_clone for upcoming 25Hz model
#281
opened Mar 30, 2026 by
ezzaldeeen
Loading…
README: replace 'huggingface-cli download' with 'hf download'
#268
opened Mar 14, 2026 by
JustACluelessKidAtSchool
Loading…
fix: prevent sample loss in chunked_decode when decoder returns fewer frames
#234
opened Feb 24, 2026 by
haosenwang1018
Loading…
fix: auto-resample audio to 24kHz instead of asserting
#233
opened Feb 24, 2026 by
haosenwang1018
Loading…
fix: preserve speaker_encoder in checkpoints to allow training resume
#232
opened Feb 24, 2026 by
haosenwang1018
Loading…
fix: preserve speaker_encoder in checkpoints and improve sample rate error
#222
opened Feb 23, 2026 by
haosenwang1018
Loading…
Add compile_codec option for 3-4x faster batch decoding
#191
opened Feb 7, 2026 by
Finrandojin
Loading…
5 tasks done
fix: apply text_projection for 600M model fine-tuning
#188
opened Feb 6, 2026 by
zhuxiaoxuhit
Loading…
Resolve each training epoch resulting in progressively faster output
#178
opened Feb 5, 2026 by
dragonfyre13
Loading…
Fix wrong project.urls (Homepage/Repository) in pyproject.toml
#152
opened Feb 1, 2026 by
MengyangGao
Loading…
Previous Next
ProTip!
Mix and match filters to narrow down what you’re looking for.