Awesome

TokShift-Transformer

This is official implementaion of paper "Token Shift Transformer for Video Classification". We achieve SOTA performance 80.40% on Kinetics-400 val. Paper link

• Updates
• Model Zoo and Baselines
• Installation
• Quick Start
• Contributors
• Citing
• Acknowledgement

Updates

July 11, 2021

• Release this V1 version (the version used in paper) to public.
• we are preparing a V2 version which include the following modifications, will release within 1 week:

1. Directly decode video mp4 file during training/evaluation
2. Change to adopt standarlize timm code-base.
3. Performances are further improved than reported in paper version (average +0.5).

April 22, 2021

• Add Train/Test guidline and Data perpariation

April 16, 2021

• Publish TokShift Transformer for video content understanding

Model Zoo and Baselines

Below is trainig log, we use 3 views evaluation (instead of 30 views) during validation for time-saving.

Installation

• PyTorch >= 1.7, torchvision
• tensorboardx

Quick Start

Train

1. Download ImageNet-22k pretrained weights from Base16 and Large16.
2. Prepare Kinetics-400 dataset organized in the following structure, trainValTest

k400
|_ frames331_train
|  |_ [category name 0]
|  |  |_ [video name 0]
|  |  |  |_ img_00001.jpg
|  |  |  |_ img_00002.jpg
|  |  |  |_ ...
|  |  |
|  |  |_ [video name 1]
|  |  |   |_ img_00001.jpg
|  |  |   |_ img_00002.jpg
|  |  |   |_ ...
|  |  |_ ...
|  |
|  |_ [category name 1]
|  |  |_ [video name 0]
|  |  |  |_ img_00001.jpg
|  |  |  |_ img_00002.jpg
|  |  |  |_ ...
|  |  |
|  |  |_ [video name 1]
|  |  |   |_ img_00001.jpg
|  |  |   |_ img_00002.jpg
|  |  |   |_ ...
|  |  |_ ...
|  |_ ...
|
|_ frames331_val
|  |_ [category name 0]
|  |  |_ [video name 0]
|  |  |  |_ img_00001.jpg
|  |  |  |_ img_00002.jpg
|  |  |  |_ ...
|  |  |
|  |  |_ [video name 1]
|  |  |   |_ img_00001.jpg
|  |  |   |_ img_00002.jpg
|  |  |   |_ ...
|  |  |_ ...
|  |
|  |_ [category name 1]
|  |  |_ [video name 0]
|  |  |  |_ img_00001.jpg
|  |  |  |_ img_00002.jpg
|  |  |  |_ ...
|  |  |
|  |  |_ [video name 1]
|  |  |   |_ img_00001.jpg
|  |  |   |_ img_00002.jpg
|  |  |   |_ ...
|  |  |_ ...
|  |_ ...
|
|_ trainValTest
   |_ train.txt
   |_ val.txt

3. Using train-script (train.sh) to train k400

#!/usr/bin/env python
import os

cmd = "python -u main_ddp_shift_v3.py \
		--multiprocessing-distributed --world-size 1 --rank 0 \
		--dist-ur tcp://127.0.0.1:23677 \
		--tune_from pretrain/ViT-L_16_Img21.npz \
		--cfg config/custom/kinetics400/k400_tokshift_div4_12x32_large_384.yml"
os.system(cmd)

Test

Using test.sh (test.sh) to evaluate k400

#!/usr/bin/env python
import os
cmd = "python -u main_ddp_shift_v3.py \
        --multiprocessing-distributed --world-size 1 --rank 0 \
        --dist-ur tcp://127.0.0.1:23677 \
        --evaluate \
        --resume model_zoo/ViT-B_16_k400_dense_cls400_segs8x32_e18_lr0.1_B21_VAL224/best_vit_B8x32x224_k400.pth \
        --cfg config/custom/kinetics400/k400_vit_8x32_224.yml"
os.system(cmd)

Contributors

VideoNet is written and maintained by Dr. Hao Zhang and Dr. Yanbin Hao.

Citing

If you find TokShift-xfmr is useful in your research, please use the following BibTeX entry for citation.

@article{tokshift2021,
  title={Token Shift Transformer for Video Classification},
  author={Hao Zhang, Yanbin Hao, Chong-Wah Ngo},
  journal={ACM Multimedia 2021},
}

Acknowledgement

Thanks for the following Github projects:

• https://github.com/rwightman/pytorch-image-models
• https://github.com/jeonsworld/ViT-pytorch
• https://github.com/mit-han-lab/temporal-shift-module
• https://github.com/amdegroot/ssd.pytorch