将Transformer引入CV届的Facebook 这次又有新发现:自监督学习 + Vision Transformers更配 自监督的Vision Transformers 自带图像语义分割的细节信息,当用于 ImageNet 分类Top 1 精度可达 80.1%! 其自监督+ViT系统DINO已开源。原地址:https://github.com/facebookresearch/dino
https://mp.weixin.qq.com/s/l5TSiuXIwOXJuLFNAmtQ8Q 源址:https://github.com/PaddlePaddle/PaddleDetection
Facebook AI 提出新型视频理解架构:完全基于Transformer的首个视频理解架构,无需卷积,训练速度快、计算成本低。 https://mp.weixin.qq.com/s/e7MWXLo9jrpVMo7lm5Q86w 原地址:https://github.com/facebookresearch/TimeSformer
PoseFormer:首个纯基于Transformer的 3D 人体姿态估计网络,性能达到 SOTA https://mp.weixin.qq.com/s/DKWSeRu_ThMf_vf9j1GCbQ 原地址:https://github.com/zczcwh/PoseFormer
大白话Pyramid Vision Transformer https://mp.weixin.qq.com/s/oJHZWmStQYYzEEOveiuQrQ 原地址:https://github.com/whai362/PVT
引燃AI社区,不用跨界也能从文本生成图像,OpenAI新模型打破自然语言与视觉次元壁 https://mp.weixin.qq.com/s/cVxFqVz_RFhAINeGC4QkTw 原地址:https://github.com/openai/dall-e
引燃AI社区,不用跨界也能从文本生成图像,OpenAI新模型打破自然语言与视觉次元壁 https://mp.weixin.qq.com/s/cVxFqVz_RFhAINeGC4QkTw 原地址:https://github.com/openai/CLIP
告别CNN?一张图等于16x16个字,计算机视觉也用上Transformer了 https://mp.weixin.qq.com/s/JKC20zRleNVIHMpt-02uag 原地址:https://github.com/google-research/vision_transformer
Transformer 又立功了!又快(420 fps)又好的车道线检测算法 https://mp.weixin.qq.com/s/hjHXWewRYh_6j5cd1J2n_w 原地址:https://github.com/liuruijin17/LSTR
屠榜各大 CV 任务的微软 Swin Transformer,近日开源了代码和预训练模型。 This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows". 原地址:https://github.com/microsoft/Swin-Transformer
MedMNIST:上海交大发布医学影像领域的MNIST https://zhuanlan.zhihu.com/p/270144930 原地址:https://github.com/MedMNIST/MedMNIST
This is the implementation of YOLObile: Real-Time Object Detection on Mobile Devices via Compression-Compilation Co-Design 原地址:https://github.com/nightsnack/YOLObile
Code for CVPR 2019 paper. BASNet: Boundary-Aware Salient Object Detection 原地址:https://github.com/xuebinqin/BASNet
The code for our newly accepted paper in Pattern Recognition 2020: "U^2-Net: Going Deeper with Nested U-Structure for Salient Object Detection." 原地址:https://github.com/xuebinqin/U-2-Net
⚡Super fast and lightweight anchor-free object detection model. 🔥Only 1.8MB and run 97FPS on cellphone🔥 原地址:https://github.com/RangiLyu/nanodet
This repository contains a list of AutoML related models and libraries. EfficientDet: Scalable and Efficient Object Detection(https://github.com/google/automl/tree/master/efficientdet) 原地址:https://github.com/google/automl
Official implementation of CVPR2020 paper "VIBE: Video Inference for Human Body Pose and Shape Estimation" 代码原地址:https://github.com/mkocabas/VIBE
This is the repository of the implemantation of cvpr 2020.
This repo holds the codes of paper: An Efficient PointLSTM for Point Clouds Based Gesture Recognition (CVPR 2020). 代码原地址:https://github.com/Blueprintf/pointlstm-gesture-recognition-pytorch