Search results

这个代码太难训练了，2025年全新架构，欢迎交流泛化384 576 768，欢迎加V交流 Mxcz2019zz

Status: Open.

#161 In primepake/wav2lip_288x288;

· sunbo11112 opened on Jan 23, 2025

坑很多，照着教程完全跑不起来，大家不要浪费时间训练了，量大找我们直接接我们api最省事 v: vip_smart_service

Status: Open.

#160 In primepake/wav2lip_288x288;

· alexhmyang opened on Dec 18, 2024

img_size384进行训练但是推理代码inference没有384的model

Status: Open.

#159 In primepake/wav2lip_288x288;

· zhang-ykun opened on Dec 2, 2024

AttributeError: module 'face_detection' has no attribute 'FaceAlignment'

Status: Open.

#158 In primepake/wav2lip_288x288;

· wxc1207 opened on Oct 14, 2024

why train_syncnet_sam.py set img_size=384?

Status: Open.

#157 In primepake/wav2lip_288x288;

· qingshuiyuyu opened on Sep 24, 2024

how to prepare training dataset

Status: Open.

#156 In primepake/wav2lip_288x288;

· dxLu opened on Sep 24, 2024

大家有遇到过执行train_syncnet_sam.py后出现的张量维度不一致问题吗：tensor a (1024) must match the size of tensor b (9216)，如图

Status: Open.

#155 In primepake/wav2lip_288x288;

· yuanmaitian opened on Sep 23, 2024

出售一份自己人工手动下载人脸数字人的数据集原始视频 ，视频分辨率1080-4k口播视频，可用于训练wav2lip-192，wav2lip-288，wav2lip-384，wav2lip-512，DINet-256，NeRF等模型。微信douzijun1999 请注明来意

Status: Open.

#154 In primepake/wav2lip_288x288;

· dousang1999 opened on Sep 15, 2024

每次训练停止，都从头开始，怎么修改能继续上次断开的地方开始

Status: Open.

#153 In primepake/wav2lip_288x288;

· zn123 opened on Sep 14, 2024

The amplitude of the mouth shape becomes smaller with custom dataset

Status: Open.

#152 In primepake/wav2lip_288x288;

· zhangziwenHIT opened on Sep 10, 2024

structure and arrangement of dataset before executing any scrip with flow of script execution!!!!!!

Status: Open.

#151 In primepake/wav2lip_288x288;

· vpssa opened on Jul 30, 2024

Use HuBERT features to train SyncNet, the loss does not converge.

Status: Open.

#150 In primepake/wav2lip_288x288;

· hnsywangxin opened on Jun 29, 2024