模型/ Model 在讨论模型前我们先思考一下, 当我们说判断音唇同步时是在指什么? 其实本质上就是在判断音频和唇形在某个共同参数空间下的相似性! Get到这一点那么模型构建的思路也就明晰了. 既然是判断相似度的问题, 那么熟悉深度学习的朋友自然而然的就明白这涉及到下面两个问题: 如何将音频和视频两个模态的特...
SyncNetModel.py new detector combined Mar 29, 2020 demo_feature.py model change Feb 22, 2019 demo_syncnet.py change data structure and video loader Feb 28, 2020 download_model.sh new detector combined Mar 29, 2020 requirements.txt new detector combined ...
python trainSyncNet.py --temporal_stride 2 --maxFrames 50 --model models.SyncNetModelFBank --save_path data/exp01 Pretrained model A pretrained model can be downloaded from here. You can check that the following script returns: EER 20.4348. python trainSyncNet.py --temporal_stride 2 --max...
def train(device, model_single, train_data_loader, test_data_loader, optimizer, checkpoint_dir=None, checkpoint_interval=None, nepochs=None): model = nn.DataParallel(model_single) global global_step, global_epoch resumed_step = global_step while global_epoch < nepochs: running_loss ...
Saved searches Use saved searches to filter your results more quickly Cancel Create saved search Sign in Sign up Reseting focus {{ message }} zhangbo2008 / wav2lip384_my2 Public Notifications You must be signed in to change notification settings Fork 7 Star 37 ...
Saved searches Use saved searches to filter your results more quickly Cancel Create saved search Sign in Sign up Reseting focus {{ message }} zhangbo2008 / wav2lip384_my2 Public Notifications You must be signed in to change notification settings Fork 8 Star 37 ...