Benchmarking the robustness of Spatial-Temporal Models

Last update: Dec 16, 2022

Overview

Benchmarking the robustness of Spatial-Temporal Models

This repositery contains the code for the paper Benchmarking the Robustness of Spatial-Temporal Models Against Corruptions.

Python 2.7 and 3.7, Pytorch 1.7+, FFmpeg are required.

Requirements

pip3 install - requirements.txt

Mini Kinetics-C

Download original Kinetics400 from link.

The Mini Kinetics-C contains half of the classes in Kinetics400. All the classes can be found in mini-kinetics-200-classes.txt.

Mini Kinetics-C Leaderboard

Corruption robustness of spatial-temporal models trained on clean Mini Kinetics and evaluated on Mini Kinetics-C.

Approach	Reference	Backbone	Input Length	Sampling Method	Clean Accuracy	mPC	rPC
TimeSformer	Gedas et al.	Transformer	32	Uniform	82.2	71.4	86.9
3D ResNet	K. Hara et al.	ResNet-50	32	Uniform	73.0	59.2	81.1
I3D	J. Carreira et al.	InceptionV1	32	Uniform	70.5	57.7	81.8
SlowFast 8x4	C. Feichtenhofer at al.	ResNet-50	32	Uniform	69.2	54.3	78.5
3D ResNet	K. Hara et al.	ResNet-18	32	Uniform	66.2	53.3	80.5
TAM	Q.Fan et al.	ResNet-50	32	Uniform	66.9	50.8	75.9
X3D-M	C. Feichtenhofer	ResNet-50	32	Uniform	62.6	48.6	77.6

For fair comparison, it is recommended to submit the result of approach which follows the following settings: Backbone of ResNet-50, Input Length of 32, Uniform Sampling at Clip Level. Any result on our benchmark can be submitted via pull request.

Mini SSV2-C

Download original Something-Something-V2 datset from link.

The Mini SSV2-C contains half of the classes in Something-Something-V2. All the classes can be found in mini-ssv2-87-classes.txt.

Mini SSV2-C Leaderboard

Corruption robustness of spatial-temporal models trained on clean Mini SSV2 and evaluated on Mini SSV2-C.

Approach	Reference	Backbone	Input Length	Sampling Method	Clean Accuracy	mPC	rPC
TimeSformer	Gedas et al.	Transformer	16	Uniform	60.5	49.7	82.1
I3D	J. Carreira et al.	InceptionV1	32	Uniform	58.5	47.8	81.7
3D ResNet	K. Hara et al.	ResNet-50	32	Uniform	57.4	46.6	81.2
TAM	Q.Fan et al.	ResNet-50	32	Uniform	61.8	45.7	73.9
3D ResNet	K. Hara et al.	ResNet-18	32	Uniform	53.0	42.6	80.3
X3D-M	C. Feichtenhofer	ResNet-50	32	Uniform	49.9	40.7	81.6
SlowFast 8x4	C. Feichtenhofer at al.	ResNet-50	32	Uniform	48.7	38.4	78.8

Training and Evaluation

To help researchers reproduce the benchmark results provided in our leaderboard, we include a simple framework for training and evaluating the spatial-temporal models in the folder: benchmark_framework.

Running the code

Assume the structure of data directories is the following:

~/
  datadir/
    mini_kinetics/
      train/
        .../ (directories of class names)
          ...(hdf5 file containing video frames)
    mini_kinetics-c/
      .../ (directories of corruption names)
        .../ (directories of severity level)
          .../ (directories of class names)
            ...(hdf5 file containing video frames)

Train I3D on the Mini Kinetics dataset with 4 GPUs and 16 CPU threads (for data loading). The input lenght is 32, the batch size is 32 and learning rate is 0.01.

python3 train.py --threed_data --dataset mini_kinetics400 --frames_per_group 1 --groups 32 --logdir snapshots/ \
--lr 0.01 --backbone_net i3d -b 32 -j 16 --cuda 0,1,2,3

Test I3D on the Mini Kinetics-C dataset (pretrained model is loaded)

python3 test_corruption.py --threed_data --dataset mini_kinetics400 --frames_per_group 1 --groups 32 --logdir snapshots/ \
--pretrained snapshots/mini_kinetics400-rgb-i3d_v2-ts-max-f32-cosine-bs32-e50-v1/model_best.pth.tar --backbone_net i3d -b 32 -j 16 -e --cuda 0,1,2,3

Benchmarking the robustness of Spatial-Temporal Models

Related tags

Overview

Benchmarking the robustness of Spatial-Temporal Models

Requirements

Mini Kinetics-C

Mini Kinetics-C Leaderboard

Mini SSV2-C

Mini SSV2-C Leaderboard

Training and Evaluation

Running the code

Owner

Yi Chenyu Ian

An implementation of EWC with PyTorch

Fast, general, and tested differentiable structured prediction in PyTorch

Source code for the NeurIPS 2021 paper "On the Second-order Convergence Properties of Random Search Methods"

For holding anime-related object classification and detection models

Learning with Subset Stacking

Pytorch implementation of FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks

Repository for "Exploring Sparsity in Image Super-Resolution for Efficient Inference", CVPR 2021

Minimalistic PyTorch training loop

[CVPR'21] Multi-Modal Fusion Transformer for End-to-End Autonomous Driving

Stochastic Tensor Optimization for Robot Motion - A GPU Robot Motion Toolkit

Development of IP code based on VIPs and AADM

Mesh TensorFlow: Model Parallelism Made Easier

official code for dynamic convolution decomposition

Official implementation of ACTION-Net: Multipath Excitation for Action Recognition (CVPR'21).

implement of SwiftNet:Real-time Video Object Segmentation

Simple data balancing baselines for worst-group-accuracy benchmarks.

Ensembling Off-the-shelf Models for GAN Training

The author's officially unofficial PyTorch BigGAN implementation.

Using a Seq2Seq RNN architecture via TensorFlow to predict future Bitcoin prices

The easiest tool for extracting radiomics features and training ML models on them.