search

mit-han-lab / temporal-shift-module

[ICCV 2019] TSM: Temporal Shift Module for Efficient Video Understanding
https://arxiv.org/abs/1811.08383
MIT License
2.07k stars 417 forks source link

Still don't know how to get val_videofolder.txt #68

Open zjzhengyin opened 4 years ago

zjzhengyin commented 4 years ago

I'm sorry. Although I've looked at other people's questions, I still wonder how to get val_videofolder.txt

Andy1621 commented 4 years ago

Could you please let us know which dataset you use? I have tried to train in Sth-Sth v1 dataset, and the val_videofolder.txt is got by run gen_label_sthv1.py, which uses validation.csv. By the way, validation.csv can be downloaded in https://20bn.com/datasets/something-something/v1

HaoyuWang00 commented 2 years ago

Could you please let us know which dataset you use? I have tried to train in Sth-Sth v1 dataset, and the val_videofolder.txt is got by run gen_label_sthv1.py, which uses validation.csv. By the way, validation.csv can be downloaded in https://20bn.com/datasets/something-something/v1

Hi @Andy1621 , I found this link no longer available. I am trying to play with kinetics. Could you plz provide a link for downloading those .csv files including kinetics_val.csv? Thanks!

HaoyuWang00 commented 2 years ago

Could you please let us know which dataset you use? I have tried to train in Sth-Sth v1 dataset, and the val_videofolder.txt is got by run gen_label_sthv1.py, which uses validation.csv. By the way, validation.csv can be downloaded in https://20bn.com/datasets/something-something/v1

Could you please let us know which dataset you use? I have tried to train in Sth-Sth v1 dataset, and the val_videofolder.txt is got by run gen_label_sthv1.py, which uses validation.csv. By the way, validation.csv can be downloaded in https://20bn.com/datasets/something-something/v1

Hi @Andy1621 , I found this link no longer available. I am trying to play with kinetics. Could you plz provide a link for downloading those .csv files including kinetics_val.csv? Thanks!

I just wonder if I need to download the dataset locally... I am trying to use Kinetics, and by the Data Preparation here, I need to generate val_videofolder.txt by running gen_label_dataset.py. Now the question becomes, what should be the input files here (in the image)? image

I went to Kinetics website and downloaded those files: image

But when I use them as input, I got an KeyError. So am I using the wrong file or using them wrongly? If it is the former case, do I need to download the dataset in video format and strictly follow the Data Preparation?

Thanks so much! I am an undergraduate student in Engineering and have not much idea in computer science/AI. I am sorry if the answer to my question is obvious. Your help will be highly appreciated!

Andy1621 commented 2 years ago

For Kinetics, you can download it in cvfoundation. By the way, welcome to play with our UniFormer, which is a new SOTA backbone for video classification.