Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/paddlepaddle/paddlevideo

Awesome video understanding toolkits based on PaddlePaddle. It supports video data annotation tools, lightweight RGB and skeleton based action recognition model, practical applications for video tagging and sport action detection.
https://github.com/paddlepaddle/paddlevideo

actbert action-detection action-localization action-recognition activitynet ava bmn kinetics400 pp-tsm slowfast st-gcn t2vlad temporal-action-detection tsm tsn video-recognition video-understanding videotag youtube-8m

Last synced: 6 days ago
JSON representation

Host: GitHub
URL: https://github.com/paddlepaddle/paddlevideo
Owner: PaddlePaddle
License: apache-2.0
Created: 2020-11-12T11:40:03.000Z (about 4 years ago)
Default Branch: develop
Last Pushed: 2024-11-23T13:45:04.000Z (about 1 month ago)
Last Synced: 2024-11-23T14:29:19.363Z (about 1 month ago)
Topics: actbert, action-detection, action-localization, action-recognition, activitynet, ava, bmn, kinetics400, pp-tsm, slowfast, st-gcn, t2vlad, temporal-action-detection, tsm, tsn, video-recognition, video-understanding, videotag, youtube-8m
Language: Python
Homepage:
Size: 106 MB
Stars: 1,534
Watchers: 38
Forks: 381
Open Issues: 182
Metadata Files:
- Readme: README.md
- Contributing: docs/CONTRIBUTING.md
- License: LICENSE

Awesome Lists containing this project

README

        [English](README_en.md) | 中文

# PaddleVideo

![python version](https://img.shields.io/badge/python-3.7+-orange.svg) ![paddle version](https://img.shields.io/badge/PaddlePaddle-2.3.1-blue)

## 简介

PaddleVideo旨在打造一套丰富、领先且实用的Video工具库，旨在帮助开发者更好的进行视频领域的学术研究和产业实践。



  




## 近期更新

- 开源视频标注工具🌟[BILS](./docs/zh-CN/annotation_tools.md)，欢迎下载安装包体验～

- 发布轻量化行为识别模型**🔥[PP-TSMv2](./docs/zh-CN/model_zoo/recognition/pp-tsm_v2.md)**, Kinetics-400精度75.16%，25fps的10s视频cpu推理时间仅需456ms.各模型性能对比[benchmark](./docs/zh-CN/benchmark.md).

- 新增[知识蒸馏](./docs/zh-CN/distillation.md)功能.

- 新增基于transformer的行为识别模型[TokenShift](https://github.com/PaddlePaddle/PaddleVideo/blob/develop/docs/zh-CN/model_zoo/recognition/tokenshift_transformer.md).

- 新增基于骨骼点的行为识别模型[2s-ACGN](https://github.com/PaddlePaddle/PaddleVideo/blob/develop/docs/zh-CN/model_zoo/recognition/agcn2s.md)、[CTR-GCN](./docs/zh-CN/model_zoo/recognition/ctrgcn.md).

- 新增单阶段时空动作检测模型[YOWO](./docs/zh-CN/model_zoo/localization/yowo.md).

👀 🌟  **《产业级视频技术与应用案例》系列课程回放链接**:  https://aistudio.baidu.com/aistudio/course/introduce/6742 🌟

																	  💖 **欢迎大家扫码入群讨论** 💖



  



- 添加成功后回复【视频】加入交流群

## 特性

支持多种Video相关前沿算法，在此基础上打造产业级特色模型[PP-TSM](docs/zh-CN/model_zoo/recognition/pp-tsm.md)和[PP-TSMv2](docs/zh-CN/model_zoo/recognition/pp-tsm_v2.md)，并打通数据生产、模型训练、压缩、预测部署全流程。



    



## 快速开始

- 一行命令快速使用: [快速开始](./docs/zh-CN/quick_start.md)

## 场景应用

PaddleVideo场景应用覆盖体育、互联网、工业、医疗行业，在PP-TSM的基础能力之上，以案例的形式展示利用场景数据微调、模型优化方法、数据增广等内容，为开发者实际落地提供示范与启发。详情可查看[应用](./applications/)。

## 文档教程

- [快速开始](./docs/zh-CN/quick_start.md)

- [安装说明](./docs/zh-CN/install.md)

- [训练/测试/推理全流程使用指南](./docs/zh-CN/usage.md)

- [PP-TSM行为识别🔥](./docs/zh-CN/model_zoo/recognition/pp-tsm.md)

  - [模型库](./docs/zh-CN/model_zoo/recognition/pp-tsm.md#7)

  - [模型训练](./docs/zh-CN/model_zoo/recognition/pp-tsm.md#4)

  - [模型压缩](./deploy/slim/)

      - [模型量化](./deploy/slim/readme.md)

      - [知识蒸馏](./docs/zh-CN/distillation.md)

  - [推理部署](./deploy/)

      - [基于Python预测引擎推理](./docs/zh-CN/model_zoo/recognition/pp-tsm.md#62)

      - [基于C++预测引擎推理](./deploy/cpp_infer/readme.md)

      - [服务端部署](./deploy/python_serving/readme.md)

      - [Paddle2ONNX模型转化与预测](./deploy/paddle2onnx/readme.md)

      - [Benchmark](./docs/zh-CN/benchmark.md)

- [前沿算法与模型](./docs/zh-CN/model_zoo/README.md)🚀

- [数据集](./docs/zh-CN/dataset/README.md)

- [场景应用](./applications/README.md)

- [数据标注](./docs/zh-CN/annotation_tools.md)

- [赛事支持](./docs/zh-CN/competition.md)

- [贡献代码](./docs/zh-CN/contribute/README.md)

## 许可证书

本项目的发布受[Apache 2.0 license](LICENSE)许可认证。