Thank you for open-sourcing such a great project. I read your training method carefully follow https://github.com/Rudrabha/Wav2Lip#training-on-datasets-other-than-lrs2. first, I trained the expert discriminator for my own dataset before training Wav2Lip. but the loss have not declined The total le...

https://github.com/xk-huang/Wav2Lip/tree/update_dockerfile

according to the information in lipsync_expert.pth, I find this model has been trained 2450 epoch, is it right ?

Dear author, Thanks for sharing the excellent work. I found that when using my personal video, there is a clear box region around the mouth in the output result, see as below: What could be the reason of this, and could you please give me some instruction on how to solve it? Many thanks for the hel...

Hi! I am trying to train the model using LRS2 dataset. The steps I followed are Download the dataset Concatenate the files Extract it to the directory I pasted the (train, val, test) .txt from lrs2 in filelists directory. preprocess it using "python preprocess.py --data_root data_root/main --prepro...

when I use a music as audio input, when the music starts, the mouth shape of the character starts to move, but the singing part has not yet reached. Is there a better solution to this problem?

How much is the maximum video size? and isn't it available for high quality video?

I continue train on LR2 dataset with pretrain model " wav2lip_gan.pth", but the result is the human's mouth don't move. Here is my sample. And here is my log. the hparams is the repo's hparmas ,I didn't change it. What cause this ? I'm so confuse about it. Load checkpoint from: /root/autodl-tmp/Wav...

Hi, I have tried the library its an exceptional work. I am doing inference but facing an issue regarding the output. The output generated have some boxy effect around the mouth. Is there any way to remove it ?

Hi, the default image merging is to DIVX video which I guess was chosen for compatibility, but it's awkward and it seems the default low quality of opencv can't be adjusted, it makes it very blocky. For that reason I made little editions in a fork: https://github.com/Twenkid/Wav2Lip-DFL-SAEHDBW/ Now...