Awesome MegEngine

• MegEngine_CU11: 包含 CUDA11 支持的 MegEngine Python Wheel 包
• MgeEditing：基于 MegEngine 的图像库
• Echo：一个优秀的算子库
• ClearML(Allegro Trains): ML/DL 开发和生产套件，包含 Experiment Manager / ML-Ops / Data-Management 三大核心模块
• MegEngine.js：MegEngine javascript 版本，可以在 javascript 环境中快速部署 MegEngine 模型
• commitlint config megengine：MegEngine commitlint 配置项
• Cascade RCNN ：一种 Cascade RCNN 实现
• Megvision：一些经典 CV 模型的实现和权重
• Models：MegEngine 实例训练代码，包含各类任务的基本实现，是初学者最好的参考物
• MgeConvert：MegEngine Traced module 模型格式转换器，支持 ONNX / TFLite / Caffe 等各类导出格式
• MegFile：一个可以完美抽象 S3、HTTP、本地文件等协议的 python 文件接口库，是 smart-open 库的升级版
• MegFlow：面向计算机视觉应用的流式计算框架，提供了一套可快速完成 AI 应用部署的视觉解析服务方案
• MegSpot：一款提供免费免登录、高效、专业、跨平台的图片&视频的对比的 PC 应用工具
• BaseCls：极其全面的分类模型库，提供了海量模型的训练代码和预训练权重，全部算法均可以快速部署到硬件上
• Swin Transformer：Swin Transformer 的实现，支持 DTR 功能减少显存占用量
• MegFLow-Cat_Feeder：基于 MegFlow 框架的猫咪检测以及自动投喂的解决方案
• MegPeak：处理器性能测试工具
• Netron 已可视化 MegEngine TracedModule，欢迎使用 模型示例文件 体验
• MegCC：一个运行时超轻量，高效，移植简单的深度学习模型编译器
• BaseDet：提供了一些经典的检测 SOTA 模型以及相关组件
  • BaseDet 使用示例
• mperf：一个算子性能调优工具箱
• MegEngine 原创论文复现：
  • RepLKNet：[CVPR 2022] Scaling Up Your Kernels to 31x31: Revisiting Large Kernel Design in CNNs
  • GyroFlow: [ICCV 2021] GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning
  • ICD: [NeurIPS 2021] Instance-Conditional Knowledge Distillation for Object Detection
  • FINet: [AAAI 2022] FINet: Dual Branches Feature Interaction for Partial-to-Partial Point Cloud Registration
  • OMNet: [ICCV 2021] OMNet: Learning Overlapping Mask for Partial-to-Partial Point Cloud Registration
  • CREStereo: [CVPR 2022] Practical Stereo Matching via Cascaded Recurrent Network with Adaptive Correlation
  • D2C-SR: [ECCV 2022] D2C-SR: A Divergence to Convergence Approach for Real-World Image Super-Resolution
  • FST-Matching: [ECCV 2022] Explaining Deepfake Detection by Analysing Image Matching
  • zipfls: [ECCV 2022] Zipf's LS: Efficient One Pass Self-distillation with Zipf's Label Smoothing
  • HDR-Transformer: [ECCV 2022] Ghost-free High Dynamic Range Imaging with Context-aware Transformer
• 经典论文模型对应结构的 MegEngine inference 函数：
  • Masked Autoencoders Are Scalable Vision Learners
  • Wide Residual Networks
  • ResNeSt: Split-Attention Networks
  • Visual Attention Network
  • A ConvNet for the 2020s
  • UniFormer: Unifying Convolution and Self-attention for Visual Recognition
  • Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection
  • Probabilistic Anchor Assignment with IoU Prediction for Object Detection
  • OTA: Optimal Transport Assignment for Object Detection
  • Pyramid Scene Parsing Network
  • GhostNet: More Features from Cheap Operations
  • Squeeze-and-Excitation Networks
  • OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
  • Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net
  • HarDNet: A Low Memory Traffic Network
  • Learning to Navigate for Fine-grained Classification
  • FishNet: a versatile backbone for image, region, and pixel level prediction
  • A Light CNN for Deep Face Representation with Noisy Labels
  • Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
  • Towards Compact Single Image Super-Resolution via Contrastive Self-distillation
  • ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks
  • Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Images
  • Densely Connected Convolutional Networks
  • Aggregated Residual Transformations for Deep Neural Networks
  • Learning Transferable Visual Models From Natural Language Supervision
  • Image Super-Resolution Using Very Deep Residual Channel Attention Networks
  • Residual Dense Network for Image Super-Resolution
  • BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation