DCNv2 supports decent pytorch 1.11

ncnn impl of ppyolov2 and ppyolo and ppyoloe.

PyTorch Deformable Convolutional Networks v2 (no compile required)

A fully cuda implementation of DCNv2(deformable convolution) forward. Without dependent of cuTorch(THC).

DCNv2_torch1.11

This is a real time instance segmentation task implemented with YOLACT++ and DCNv2 on Google Colab.

Some C++/C/CUDA Extension

CenterNet Fixed For Google Colab

Useful plugins for Pytorch1.5+ and TensorRT7/8.

[Pattern Recognition]Video Saliency Prediction using Enhanced Spatiotemporal Alignment Network

NN-based radar-camera post sensor fusion implemented by TensorRT

Don't feel pain to use Deformable Convolution

More readable and flexible yolov5 with more backbone(gcn, resnet, shufflenet, moblienet, efficientnet, hrnet, swin-transformer, etc) and (cbam，dcn and so on), and tensorrt

练习下用pytorch来复现下经典的推荐系统模型, 如MF, FM, DeepConn, MMOE, PLE, DeepFM, NFM, DCN, AFM, AutoInt, ONN, FiBiNET, DCN-v2, AFN, DCAP等

DCNv2 supports decent pytorch such as torch 1.5+ (now 1.8+)