Representative Library for video understanding
- facebookresearch/PyTorchVideo [github]
- open-mmlab/MMAction2 [github]
- paddlepaddle/PaddleVideo [github]
Compare Libraries
PyTorchVideo |
MMAction2 |
PaddleVideo |
|
Support
|
- Slow & SlowFast - X3D - MViT |
- Slow & SlowFast - X3D - TimeSformer - BMN - PoseC3D |
- PP-TSM - AGCN - TimesFormer - BMN - AttentianLSTM |
| Feature | - Support Accelerator for optimizing target device - Using object detector from Detectron2 |
- Support 4 video understanding tasks - Modular design |
- Support 3 video understanding tasks - Support Faster training stratege |
'Study > AI' 카테고리의 다른 글
| 데이터가 충분하다고 말하려면 얼마나 있어야 할까? (0) | 2022.01.24 |
|---|---|
| 딥러닝이란 무엇일까? (0) | 2022.01.15 |
| Knowledge Distillation 구현 (0) | 2021.11.29 |
| [Pytorch] Tips for Loading Pre-trained Model (0) | 2021.11.27 |
| Lightweight Deep Learning (0) | 2021.11.24 |