EmmiSpace / src /f5_tts /eval /eval_infer_batch.py

Commit History

Add flash_attn2 support attn_mask, minor fixes (#1066)
1d923b1

Zhikang Niu SWivid commited on

formatting, sorting
37cdbe8

SWivid commited on

pre-commit update and formatting
4a7cbb5

SWivid commited on

1.0.0 F5-TTS v1 base model with better training and inference performance
7ccc021

SWivid commited on

support command line set args
4dd981f

zkniu commited on

fix vocoder generate sample and #467
ee3fb21

zkniu commited on

minor fix. speech_edit & eval_infer_batch
e1470ef

SWivid commited on

minor fix
4b2c6ea

SWivid commited on

refactor: del global params and set vocos as default vocoder, add dtype check
eb1b260

zkniu commited on

refactor: more details about bigvgan, clear function definition
0f2bf40

zkniu commited on

change some infer function to support two vocoder
e54a793

zkniu commited on

update Bigvgan vocoder and F5-bigvgan version, trained on Emilia ZH&EN, 1.25m updates
851eb76

zkniu commited on

finish eval dependencies; update infer_gradio with chat feature
e3011f1

SWivid commited on

update finetune-cli -gradio
3a3f3ea

SWivid commited on

make a structure first
23a1101

SWivid commited on