-
Notifications
You must be signed in to change notification settings - Fork 2.5k
Pull requests: FunAudioLLM/CosyVoice
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Add CUDA 12.8 Docker runtime with pinned dependencies
#1901
opened May 30, 2026 by
LingLambda
Loading…
perf(inference): streaming TRT optimizations and optional bucketed Flow engines
#1894
opened May 22, 2026 by
BeckYang26
Loading…
6 tasks
feat: support pretrained_to_huggingface functionality for CosyVoice3 RL trainging
#1890
opened May 18, 2026 by
Sakkana
Loading…
fix: read wav via soundfile to avoid torchcodec dependency
#1887
opened May 18, 2026 by
jasagiri
Loading…
1 task
feat: add Apple Silicon (MPS) support for macOS ARM64
#1869
opened Apr 5, 2026 by
jasagiri
Loading…
6 of 7 tasks
Improve inference consistency and runtime stability for streaming and vLLM backends
#1864
opened Mar 25, 2026 by
KEVINTUAN12
Loading…
fix: AttributeError when spks=None in ConditionalCFM.solve_euler (issue #1854)
#1856
opened Mar 18, 2026 by
huyyxyshare
Loading…
fix: DPO implementation bug in forward_dpo (issue #1449)
#1855
opened Mar 18, 2026 by
huyyxyshare
Loading…
fix: use local token_hop_len in streaming loop to avoid concurrent st…
#1849
opened Mar 12, 2026 by
Caxson
Loading…
fix: resolve tensor shape mismatch and index out-of-bounds in CausalM…
#1848
opened Mar 12, 2026 by
Caxson
Loading…
Remove openai-whisper dependency for log_mel_spectrogram
#1846
opened Mar 12, 2026 by
musselmanjoey
Loading…
fix: support batched inference in solve_euler by removing hardcoded batch size
#1827
opened Feb 11, 2026 by
Mr-Neutr0n
Loading…
3 tasks
fix: avoid variable shadowing of timestep
t in compute_loss
#1826
opened Feb 11, 2026 by
Mr-Neutr0n
Loading…
2 tasks done
Update webui.py with consistent model name of CosyVoice3-0.5B and add installation for AMD ROCm GPU
#1754
opened Dec 28, 2025 by
alexhegit
Loading…
feat: add OpenAI-compatible streaming server with WebSocket support
#1751
opened Dec 27, 2025 by
coding-crying
Loading…
Webui Update model directory and modify default prompt text
#1737
opened Dec 25, 2025 by
BruceLee569
Loading…
add test code for zero-shot, sft and speaker_id testing
#1668
opened Dec 14, 2025 by
petercai
Loading…
在语音合成和自然语言控制中重新使用保存的3秒急速复刻音色。expand to reuse the saved zero_shot_spk_id in sft, instruct2
#1475
opened Jul 23, 2025 by
WLimin
Loading…
Previous Next
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.