tuna2134/sbv2-v2

Go to file

Lvjinhong 31de84eb3b 更新并完善分布式训练功能 (#186 )

* 修复多机训练问题

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* 更新并完善分布式训练功能

近期融合V2版本代码时发现之前修改的多机功能并不正确，仍会报错，只不过单机多卡情况下local_rank即相当于rank，感知不出
1. 修复train_ms.py中DDP初始化及.cuda绑定到local_rank上
2. 在default_config.yml配置文件中添加env变量 LOCAL_RANK，否则默认情况下会key error
3. 添加run_MnodesAndMgpus.sh，更新分布式相关说明

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

2023-11-16 13:46:15 +08:00

.github/workflows

feat: 增加自动格式化代码的 action

2023-09-06 21:29:44 +08:00

sync emo branch (#159 )

2023-11-08 17:09:15 +08:00

Update config.json

2023-10-30 22:28:37 +08:00

emotional/wav2vec2-large-robust-12-ft-emotion-msp-dim

Dev no emo (#123 )

2023-10-28 10:36:21 +08:00

Update esd.list

2023-09-05 21:59:10 +08:00

Dev no emo (#123 )

2023-10-28 10:36:21 +08:00

monotonic_align

Apply Code Formatter Change

2023-09-06 13:42:05 +00:00

Fix: Add missing torch.cuda.empty_cache() (#161 )

2023-11-08 19:34:48 +08:00

Fix japanese.py (#151 )

2023-11-06 06:41:41 +08:00

use diy loguru (#164 )

2023-11-09 17:38:15 +08:00

.gitignore

修复多机训练问题 (#143 )

2023-11-02 10:38:45 +08:00

.gitmodules

Dev no emo (#123 )

2023-10-28 10:36:21 +08:00

.pre-commit-config.yaml

[pre-commit.ci] pre-commit autoupdate (#180 )

2023-11-14 08:44:49 +08:00

attentions_onnx.py

Onnx导出 (#176 )

2023-11-11 18:24:03 +08:00

attentions.py

Format & polish training and webui

2023-09-12 02:21:48 -04:00

bert_gen.py

Auto download missing model for bert_gen.py (#146 )

2023-11-04 04:06:10 +08:00

commons.py

Dev no emo (#123 )

2023-10-28 10:36:21 +08:00

config.py

Add language identification. (#153 )

2023-11-06 21:56:09 +08:00

data_utils.py

use diy loguru (#164 )

2023-11-09 17:38:15 +08:00

default_config.yml

更新并完善分布式训练功能 (#186 )

2023-11-16 13:46:15 +08:00

emo_gen.py

Create emo_gen.py

2023-11-07 12:58:39 +08:00

export_onnx.py

Onnx导出 (#176 )

2023-11-11 18:24:03 +08:00

infer.py

Support multilang generation (#185 )

2023-11-15 20:22:48 +08:00

LICENSE

更改为AGPL-V3

2023-09-02 00:14:02 +08:00

losses.py

Format & polish training and webui

2023-09-12 02:21:48 -04:00

mel_processing.py

Update mel_processing.py

2023-11-05 12:25:37 +08:00

models_onnx.py

Onnx导出 (#176 )

2023-11-11 18:24:03 +08:00

models.py

更新 models.py

2023-11-10 13:54:39 +08:00

modules.py

Format & polish training and webui

2023-09-12 02:21:48 -04:00

preprocess_text.py

Fix incorrect config train_path and val_path (#125 )

2023-10-28 12:19:01 +08:00

re_matching.py

Dev no emo (#123 )

2023-10-28 10:36:21 +08:00

README.md

Update README.md

2023-11-09 20:07:29 +08:00

requirements.txt

Add language identification. (#153 )

2023-11-06 21:56:09 +08:00

resample.py

Dev no emo (#123 )

2023-10-28 10:36:21 +08:00

run_MnodesAndMgpus.sh

更新并完善分布式训练功能 (#186 )

2023-11-16 13:46:15 +08:00

server_fastapi.py

Fix: roll back #170 and fix bug in api /tools/random_example (#177 )

2023-11-11 20:28:31 +08:00

server.py

Fix: roll back #170 and fix bug in api /tools/random_example (#177 )

2023-11-11 20:28:31 +08:00

train_ms.py

更新并完善分布式训练功能 (#186 )

2023-11-16 13:46:15 +08:00

transforms.py

Apply Code Formatter Change

2023-09-06 13:42:05 +00:00

update_status.py

Dev no emo (#123 )

2023-10-28 10:36:21 +08:00

utils.py

Add model download to train_ms.py (#150 )

2023-11-06 06:42:07 +08:00

webui.py

Support multilang generation (#185 )

2023-11-15 20:22:48 +08:00

README.md

Bert-VITS2

VITS2 Backbone with multipal-language bert

请注意，本项目核心思路来源于anyvoiceai/MassTTS 一个非常好的tts项目

MassTTS的演示demo为ai版峰哥锐评峰哥本人,并找回了在金三角失落的腰子

成熟的旅行者/开拓者/舰长/博士/sensei/猎魔人/喵喵露/V应当参阅代码自己学习如何训练。

严禁将此项目用于一切违反《中华人民共和国宪法》，《中华人民共和国刑法》，《中华人民共和国治安管理处罚法》和《中华人民共和国民法典》之用途。

严禁用于任何政治相关用途。

Video:https://www.bilibili.com/video/BV1hp4y1K78E

Demo:https://www.bilibili.com/video/BV1TF411k78w

References

感谢所有贡献者作出的努力