All issues

Search results

发现训练出来的模型维度和直接下载的不一样，重新训练的是70维，直接下载官方的是73维度，而且效果比官方下载的差很多，想问下是数据处理和训练代码没有更新吗，可以具体告诉下73维度的训练细节吗

Status: Open.

#60 In jdh-algo/JoyVASA;

· zhanlv123 opened on Apr 25, 2026

The motion generation model with wav2vec2 encoder will be supported later.

Status: Open.

#59 In jdh-algo/JoyVASA;

· jaker-uu opened on Apr 15, 2026

Critical bug in the implementation of the PositionalEncoding class

Status: Open.

#58 In jdh-algo/JoyVASA;

· wvinzh opened on Dec 3, 2025

评估代码可否公开

Status: Open.

#57 In jdh-algo/JoyVASA;

· dttttttyfyfyf opened on Nov 5, 2025

dataset

Status: Open.

#56 In jdh-algo/JoyVASA;

· AG550 opened on Sep 7, 2025

SyncNet and QAlign

Status: Open.

#55 In jdh-algo/JoyVASA;

· wangli000 opened on Jul 25, 2025

可以支援4K半身、横向16：9图片吗？

Status: Open.

#54 In jdh-algo/JoyVASA;

· hzc12321 opened on Jul 14, 2025

shape of motion_feat_dim

Status: Open.

#53 In jdh-algo/JoyVASA;

· passenger1366 opened on Jul 8, 2025

现在输入输出的尺寸必须是1：1吗，这个大小能进行修改吗

Status: Open.

#52 In jdh-algo/JoyVASA;

· Gltton opened on May 27, 2025

audio length invalid! audio: 98707, coef: 128000.0

Status: Open.

#51 In jdh-algo/JoyVASA;

· QzZhang3 opened on May 6, 2025

training diffusion step = 50？

Status: Open.

#50 In jdh-algo/JoyVASA;

· YingjunShang opened on Apr 30, 2025

动物嘴巴很奇怪

Status: Open.

#49 In jdh-algo/JoyVASA;

· 972461099 opened on Apr 14, 2025