Issues

• License terms for the LatentSync 1.6 model weights

  bytedance/LatentSync#371 · Filter by author GylleneP opened on Oct 25, 2026

• License terms for LatentSync 1.5 in hosted and on-premises products

  bytedance/LatentSync#370 · Filter by author atimisru-stack opened on Oct 21, 2026

• Clarification of commercial usage rights for LatentSync 1.6 pretrained weights

  bytedance/LatentSync#369 · Filter by author milickiydenis-dev opened on Oct 20, 2026

• TensorRT-RTX optimized inference path and reduced-step speed benchmarks

  bytedance/LatentSync#365 · Filter by author PetrusVermaak opened on Jul 3, 2026

• 因为pytorch版本太低，无法在50系显卡上运行

  bytedance/LatentSync#362 · Filter by author haohetao opened on Apr 29, 2026

• 下牙不能看啊，一排小碎牙，哈哈 啥原因呢？

  bytedance/LatentSync#356 · Filter by author aireelmagic opened on Feb 3, 2026
  4

• Please consider to update to make it work with Nvidia RTX 5090

  bytedance/LatentSync#355 · Filter by author Santelomar opened on Jan 22, 2026

• 为何我的生成结果一直是脸部红色，嘴部模糊状

  bytedance/LatentSync#353 · Filter by author zjm8621491 opened on Jan 11, 2026

• Which face detection model was used for the v1.5 model and v1.6 models?

  bytedance/LatentSync#352 · Filter by author paulgeorge opened on Jan 8, 2026

• hi,跑出来之后脸部出现了脱框是怎么回事

  bytedance/LatentSync#350 · Filter by author gangxu822 opened on Dec 10, 2025
  2

• evaluation set and fvd score

  bytedance/LatentSync#349 · Filter by author jikyong opened on Dec 10, 2025

• 请问这个项目后续还会更新吗？

  bytedance/LatentSync#346 · Filter by author tianlinzx opened on Nov 25, 2025

• 怎么出来的是这样效果？请教一下哪里配置问题？

  bytedance/LatentSync#345 · Filter by author 1374232323 opened on Oct 30, 2025
  7

• 都是非实时， wan2.2-s2v和InfiniteTalk一出来，当前项目存在还有意义吗？

  bytedance/LatentSync#344 · Filter by author cacard opened on Oct 22, 2025
  5

• 4090 24g显存 使用1.6版本，输入18s的视频和音频，推演时间需要2分多钟正常吗？还有什么方式可以提升速度的呢？

  bytedance/LatentSync#343 · Filter by author XianCH opened on Oct 17, 2025

• 关于256*256的视频与512*512的视频是否共用一个syncnet

  bytedance/LatentSync#339 · Filter by author myzap opened on Sep 25, 2025
  2

• 这里有更快的高清数字人sdk，欢迎测试👏

  bytedance/LatentSync#337 · Filter by author oneCodeSuperman opened on Sep 23, 2025
  4

• Glitches instead mouth

  bytedance/LatentSync#336 · Filter by author pitkevich80 opened on Sep 21, 2025
  2

• 因為顯存不夠，只有16GB，想詢問是否能使用v1.5的版本？

  bytedance/LatentSync#335 · Filter by author HaoHaohedgehog opened on Sep 15, 2025
  1

• hello，重训Unet后infer画质会下降，人脸有一点抖动，这个怎么解决

  bytedance/LatentSync#334 · Filter by author gangxu822 opened on Sep 9, 2025
  3

• hello，训练syncnet一般需要多少数据？训练时候也不是从0开始训吧，要把如下的resume_ckpt_path设置成作者发布的pretrain model（checkpoints/stable_syncnet.pt）?

  bytedance/LatentSync#333 · Filter by author gangxu822 opened on Sep 6, 2025
  1

• Use SageAttention2?

  bytedance/LatentSync#332 · Filter by author btakita opened on Sep 5, 2025
  1

• 什么情况下需要重训syncnet网络？

  bytedance/LatentSync#331 · Filter by author gangxu822 opened on Sep 4, 2025
  12

• train_unet时候这个报错什么问题？

  bytedance/LatentSync#330 · Filter by author gangxu822 opened on Sep 4, 2025

• 请教train_unet的问题？

  bytedance/LatentSync#329 · Filter by author gangxu822 opened on Sep 4, 2025