Video Classification

Supported Models & Datasets

Action Recognition

• X3D (CVPR 2020) (Coming Soon)
• UniFormer (ICLR 2022)

Action Detection

• SlowFast (ICCV 2019) (Coming Soon)

Datasets

• Kinetics-400/600/700 (CVPR 2017)
• AVA (CVPR 2018) (Coming Soon)

Pre-trained Models & Benchmarks

Refer to MODELS.md for supported pre-trained models and benchmarks of SOTA models.

Installation

• python >= 3.6
• torch >= 1.9.0
• torchvision >= 0.10.0

Then, clone and install this repo with:

$ git clone https://github.com/sithu31296/video-classification
$ cd video-classification
$ pip install -e .

Inference

Use the following script to test the pre-trained model:

$ python tools/infer.py \
    --source VIDEO_FILE_NAME
    --model MODEL_NAME
    --model_path PRETRAINED_MODEL_PATH
    --num_classes DATASET_NUM_CLASSES

You will get the top-5 score similar to this:

Class            Score (%)
-------------  -----------
archery              91.49
throwing axe          0.11
slacklining           0.06
feeding fish          0.06
rock climbing         0.05

Training & Evaluation

Dataset Preparation

Follow the steps provided in DATASETS.md.

Configuration

Coming Soon

Training

Coming Soon

Evaluation

Coming Soon

References

• facebookresearch/SlowFast

Citations

@misc{li2022uniformer,
      title={UniFormer: Unified Transformer for Efficient Spatiotemporal Representation Learning}, 
      author={Kunchang Li and Yali Wang and Peng Gao and Guanglu Song and Yu Liu and Hongsheng Li and Yu Qiao},
      year={2022},
      eprint={2201.04676},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}