diodiogod c7cd8c8274 Version 4.13.2 ...

Fix IndexTTS-2 QwenEmotion transformers compatibility (#90)

- Change deprecated dtype parameter to torch_dtype for AutoModelForCausalLM
- Add KV-cache disabling to prevent incompatibilities with newer transformers versions
- Resolves Qwen3ForCausalLM unexpected keyword argument error

2025-10-30 00:06:35 -03:00

..

adapters

Add per-segment parameter support to streaming pipeline

2025-10-29 08:12:49 -03:00

chatterbox

Version 4.13.1

2025-10-29 23:48:11 -03:00

chatterbox_official_23lang

Version 4.13.1

2025-10-29 23:48:11 -03:00

f5_tts

Add .to() methods for F5-TTS and Higgs Audio 2 engines

2025-10-27 22:06:16 -03:00

f5tts

Version 4.10.2

2025-09-23 15:48:09 -03:00

higgs_audio

Add .to() methods for F5-TTS and Higgs Audio 2 engines

2025-10-27 22:06:16 -03:00

index_tts

Version 4.13.2

2025-10-30 00:06:35 -03:00

processors

Use consistent emoji for parameters (📊) across Index2 and VibeVoice

2025-10-29 08:05:17 -03:00

rvc

Fix RVC Clear VRAM button - register models with ComfyUI tracking

2025-10-27 23:02:23 -03:00

vibevoice_engine

Fix tokenizer caching and console cleanup

2025-10-29 01:27:14 -03:00

video

Version 4.5.25

2025-08-27 19:41:46 -03:00

__init__.py

Major refactoring: Reorganize project structure for better maintainability

2025-08-02 18:50:26 -03:00