search

Rudrabha / Wav2Lip

This repository contains the codes of "A Lip Sync Expert Is All You Need for Speech to Lip Generation In the Wild", published at ACM Multimedia 2020. For HD commercial model, please try out Sync Labs
https://synclabs.so
9.8k stars 2.13k forks source link

Invalid preliminary data set #189

Closed wowopc closed 3 years ago

wowopc commented 3 years ago

Run code

cd C:\Wav2Lip&& python preprocess.py --data_root /mvlrs_v1 --preprocessed_root /lrs2_preprocessed

result be understood as (type, (1,)) / '(1,)type'. np_resource = np.dtype([("resource", np.ubyte, 1)]) Started processing for /mvlrs_v1 with 1 GPUs 0it [00:00, ?it/s] Dumping audios... 0it [00:00, ?it/s]

Why not run?

wowopc commented 3 years ago

Run code python preprocess.py --data_root mvlrs_v1\main --preprocessed_root lrs2_preprocessed/ result IndexError: list index out of range Traceback (most recent call last): File "preprocess.py", line 105, in main process_audio_file(vfile, args) File "preprocess.py", line 71, in process_audio_file dirname = vfile.split('/')[-2] IndexError: list index out of range 100%|██████████████████████████████████████████████████████████████████████████| 48165/48165 [00:10<00:00, 4568.66it/s] Why is it wrong again...

prajwalkr commented 3 years ago

The preprocessing script only handles LRS2. Not mvlrs_v1. You need to change the code appropriately to do that.

wowopc commented 3 years ago

运行 cd C:\Wav2Lip &&preprocess.py --data_root data_root/main --preprocessed_root lrs2_preprocessed/ no effect...

2: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'. np_resource = np.dtype([("resource", np.ubyte, 1)]) Started processing for data_root/main with 1 GPUs 0it [00:00, ?it/s] Dumping audios... 0it [00:00, ?it/s]

backpass commented 3 years ago

Hi @prajwalkr

I also have downloaded LRS2 dataset (from https://www.robots.ox.ac.uk/~vgg/data/lip_reading/lrs2.html) and the folder "mvlrs_v1" has been created. Is it okay and did you get the same title after unpacking?

I am suspicious because there is an actual dataset called "MV-LRS" so it may be possible that the wrong dataset is distributed on this website.

prajwalkr commented 3 years ago

Is it okay and did you get the same title after unpacking?

Yes, and yes.

backpass commented 3 years ago

Thanks for clarification, I was a bit nervous about it )