- 09 Dec, 2025 3 commits
-
-
Musisoul authored
-
LiangLiu authored
-
Gu Shiqiao authored
-
- 08 Dec, 2025 3 commits
-
-
PengGao authored
-
LiangLiu authored
Tidy VAReader & OmniVAReader Tidy VARecorder & X264VARecorder VARecorder with stream, use buffer stream Tidy env WORKER_RANK, READER_RANK, RECORDER_RANK Support voice type choose
-
Gu Shiqiao authored
-
- 05 Dec, 2025 4 commits
-
-
XHPlus authored
-
Gu Shiqiao authored
-
Gu Shiqiao authored
-
Gu Shiqiao authored
-
- 04 Dec, 2025 11 commits
-
-
Watebear authored
-
Kane authored
1. 增加mlu dockerfile 2. 增加dockerfile保存目录
-
LiangLiu authored
deploy 部署相关环境更新
-
PengGao authored
-
Yang Yong (雍洋) authored
-
Yang Yong (雍洋) authored
-
Yang Yong (雍洋) authored
-
sandy authored
Co-authored-by:Yang Yong (雍洋) <yongyang1030@163.com>
-
helloyongyang authored
-
helloyongyang authored
-
helloyongyang authored
-
- 03 Dec, 2025 9 commits
-
-
Musisoul authored
-
PengGao authored
Co-authored-by:helloyongyang <yongyang1030@163.com>
-
Yang Yong (雍洋) authored
-
Gu Shiqiao authored
Co-authored-by:helloyongyang <yongyang1030@163.com>
-
LiangLiu authored
- 服务化功能新增(前端+后端): 1、seko-talk 模型支持多人输入 2、支持播客合成与管理 3、支持wan2.2 animate 模型 - 后端接口新增: 1、 基于火山的播客websocket合成接口,支持边合成边听 2、播客的查询管理接口 3、基于 yolo 的多人人脸检测接口 4、音频多人切分接口 - 推理代码侵入式修改 1、将 animate 相关的 输入文件路径(mask/image/pose等)从固定写死的config中移除到可变的input_info中 2、animate的预处理相关代码包装成接口供服务化使用 @xinyiqin --------- Co-authored-by:qinxinyi <qxy118045534@163.com>
-
Yang Yong (雍洋) authored
-
Musisoul authored
-
Yang Yong (雍洋) authored
-
PengGao authored
-
- 02 Dec, 2025 3 commits
-
-
yihuiwen authored
Co-authored-by:yihuiwen <yihuiwen@sensetime.com>
-
Yang Yong (雍洋) authored
-
Yang Yong (雍洋) authored
-
- 01 Dec, 2025 1 commit
-
-
PengGao authored
-
- 28 Nov, 2025 4 commits
-
-
Kane authored
1. 修复muxi int8-vllm推理结果精度问题。torch.empty导致推理结果有nan值。 Co-authored-by:root <root@master.cluster.local>
-
Gu Shiqiao authored
-
Watebear authored
-
Gu Shiqiao authored
-
- 27 Nov, 2025 2 commits
-
-
Kane authored
1. 修复mlu int8量化
-
Yang Yong (雍洋) authored
-