skhcjh231 / matr_codebase

Online Temporal Action Localization with Memory-Augmented Transformer

Youngkil Song*, Dongkeun Kim*, Minsu Cho, Suha Kwak

Requirements

• Ubuntu 22.04
• Python 3.10.9
• CUDA 11.7
• Pytorch 2.0.0

Environment installation

pip install -r requirements.txt

Download datasets & trained weights (THUMOS14 dataset only)

link

Run train scripts

sh scripts/train_ontal.sh

Run test scripts

sh scripts/eval_ontal.sh

File structure

├── data/
│   ├── thumos_all_feature_val_V3.pickle
│   ├── thumos_all_feature_test_V3.pickle
│   └── thumos14_v2.json
├── checkpoint/
│   └── best_epoch.pth
├── criterion/ 
├── Evaluation/ 
├── models/ 
├── scripts/ 
├── util/ 
dataset.py 
eval.py 
main.py 
on_tal_task.py 
requirements.txt  
readme.md