Wav2Lip
Wav2Lip copied to clipboard
Training stuck
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 --preprocessed_root lrs2_preprocessed/" command when I tried to start training the syncnet model it stuck in this point. same thing happens when I tried to run the original model training with pretrained sycnet model. The output where it stucks is below:
Kindly let me know whats the issue as I followed all the steps mentioned in the readme file.
@Rudrabha any update regarding this
@Rudrabha any comments on this i am facing the same issue
Probably it means that getitem method of dataset class fails to find / read your data
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 --preprocessed_root lrs2_preprocessed/" command when I tried to start training the syncnet model it stuck in this point. same thing happens when I tried to run the original model training with pretrained sycnet model. The output where it stucks is below:
Kindly let me know whats the issue as I followed all the steps mentioned in the readme file.
what does "Conf' mean in the .txt of the mvlrs_v1?
sample:
00001.txt Text: WHEN YOU'RE COOKING CHIPS AT HOME Conf: 4
Hi , any update on this will be helpful ! I am training on the custom dataset and facing the same issue.
Also it will be helpful , if any example what will be in train.txt file.
Thank you
any updates facing the same issue and kindly also explain the structure of root dataset and processed dataset and what must be in txt files