前言

飞桨论文复现活动采用线上赛事+课程相结合模式,该活动以顶会论文复现为主题,汇聚众多人工智能高阶开发者参与,逐渐形成了围绕学术前沿模型理论研究、SOTA模型复现、算法研究与应用为核心的生态活动。

最新一期活动涉及AI领域前沿论文百余篇,更有最新论文MAE、Swin Transformer V2等你来挑战~

飞桨论文复现挑战赛(第五期)🔥 进行中:比赛地址

飞桨论文复现活动已经连续举办4届,相关论文覆盖AI众多细分领域,包括图像分类、目标检测、图像分割、OCR、GAN、人脸识别、推荐系统、图网络、自然语言处理、强化学习等众多方向。活动累计吸引高阶开发者达3000+,开发者累计开源飞桨Github生态模型100+,论文复现AI Studio相关项目总浏览量达20000+

论文复现的价值

论文复现是深入掌握前沿论文的最佳路径,它不仅仅能够帮助开发者掌握细分方向的重要工作成果,同时能够帮助大家实现更多的价值,飞桨论文复现逐渐成为深入掌握模型实践技术的最优方式。

如何完成论文复现

无论你是科研小白,还是高阶开发者,科学的论文复现方法都会使得复现顶会论文事半功倍,具体相关经验与常见问题FAQ见官方《论文复现赛指南》

第三、四期精品项目合集

note:预览模型展现不全,可运行项目查看详情list。

🚀 目标检测

论文精品项目顶会
Scaled-YOLOv4: Scaling Cross Stage Partial Network基于飞桨复现ScaledYOLOv4CVPR 2021
Sparse R-CNN: End-to-End Object Detection with Learnable Proposals基于 PPDet 框架的 SparseR-CNN 复现CVPR 2021
EfficientDet: Scalable and Efficient Object Detection目标检测:基于Paddle框架的EfficientDet论文复现CVPR 2020
Hybrid Task Cascade for Instance SegmentationHTC 飞桨论文复现挑战赛(第四期)CVPR 2019
YOLO9000: Better, Faster, Stronger目标检测:基于飞桨复现YOLOV2CVPR 2017
Focal Loss for Dense Object Detection基于 PPDet 框架的 RetinaNet 复现ICCV 2017
You Only Look Once: Unified, Real-Time Object Detection基于PaddlePaddle框架的YOLOv1复现CVPR 2016
YOLOv4: Optimal Speed and Accuracy of Object Detection论文复现赛第四期YOLOv4冠军代码

🚀 图像分类

论文精品项目顶会
MLP-Mixer: An all-MLP Architecture for Vision基于飞桨复现MLP-MixerNeurIPS 2021
R-Drop: Regularized Dropout for Neural NetworksR-Drop:摘下SOTA的Dropout正则化策略NeurIPS 2021
Wide Residual NetworksResNet_wide for CIFAR10
Deep Networks with Stochastic DepthDeep Networks with Stochastic Depth

🚀 图像分割

论文精品项目顶会
Rethinking BiSeNet For Real-time Semantic Segmentation论文复现:基于Paddle复现Rethinking BiSeNetCVPR 2021
Exploring Cross-Image Pixel Contrast for Semantic Segmentation基于PPSeg框架的HRNet_W48_Contrast复现ICCV 2021
Real-Time High-Resolution Background Matting抠像任务:基于飞桨复现BackgroundMattingV2模型CVPR 2021
PointRend: Image Segmentation as RenderingPointRend:将图像分割视作渲染问题,显著提升语义/实例分割性能CVPR 2020
Context Prior for Scene Segmentation基于Paddleseg框架的CPNet复现CVPR 2020
ESPNetv2: A Light-weight, Power Efficient, and General Purpose Convolutional Neural Network基于Paddleseg框架的ESPnetV2复现CVPR 2019
YOLACT: Real-time Instance Segmentation【飞桨框架论文复现】YOLACT Real-time Instance SegICCV 2019
PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation基于飞桨框架复现PointNetCVPR 2017
PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space点云处理: 基于飞桨复现PointNet++NeurIPS 2017
Holistically-Nested Edge Detection论文复现赛第四期《Holistically-Nested Edge 》冠军方案ICCV 2015
SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation基于 PaddlePaddle 框架的 SegNet 论文复现
PortraitNet: Real-time portrait segmentation network for mobile device论文复现赛第三期《PortraitNet:移动设备实时人像分割网络》第1名方案
End to End Learning for Self-Driving Cars论文复现:实现简单的端到端自动驾驶模型
CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning论文复现:基于 Paddle2.0 复现 CheXNet 模型

🚀 图像生成

论文精品项目顶会
Encoding in Style: a StyleGAN Encoder for Image-to-Image Translationpixel2style2pixel第四期论文复现赛第一名代码CVPR 2021
PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models图像超分:基于飞桨复现PULSECVPR 2020
Progressive Growing of GANs for Improved Quality, Stability, and Variation论文复现:Progressive Growing of GANs(PGAN)ICLR 2018
Deep Image PriorDIP第四期论文复现赛第一名代码CVPR 2018
Generative Adversarial Text to Image Synthesis文本生成图像:基于飞桨复现Text2Imgae
Only a Matter of Style: Age Transformation Using a Style-Based Regression Model人脸生成:基于飞桨复现SAM模型
Adversarial Autoencoders对抗性自编码器:基于飞桨复现AAE

🚀 多模态

论文精品项目顶会
Vilbert: Pretraining task-agnostic visiolinguisti representations for vision-and-language tasksViLBERT-REC-PaddleNeurIPS 2019
Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering【飞桨框架论文复现】Bottom-Up and Top-DownCVPR 2018
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention【飞桨框架论文复现】Show, Attend and TellICML 2015

🚀 自然语言处理

论文精品项目顶会
ChineseBERT: Chinese Pretraining Enhanced by Glyph and Pinyin Information百度第四届论文复现赛 ChineseBERT冠军复现代码ACL 2021
Recipes for building an open-domain chatbot飞浆论文复现挑战赛 - Blenderbot复现EACL 2021
ConvBERT: Improving BERT with Span-based Dynamic Convolution[论文复现赛第4期冠军] ConvBERTNeurIPS 2020
MPNet: Masked and Permuted Pre-training for Language Understanding[论文复现赛第4期冠军] MPNetNeurIPS 2020
Reformer: The Efficient Transformer[论文复现赛第4期冠军] ReformerICLR 2020
End-To-End Memory Networks基于Paddle复现论文End to End Memory NetworksNeurIPS 2015

🚀 推荐系统

论文精品项目顶会
A Dual Input-aware Factorization Machine for CTR Prediction论文复现推荐赛道第 2 弹之基于PaddleRec 快速复现 DIFM 算法IJCAI 2020
FAT-DeepFFM: Field Attentive Deep Field-aware Factorization Machine飞浆论文复现挑战赛(第四期)Fat_DeepFFM模型第一名复现方案IJCAI 2020
Deep Match to Rank Model for Personalized Click-Through Rate Prediction[论文复现赛第3期][DMR第1名方案]DMR with PaddleRecAAAI 2020
Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding【论文复现赛第4期】Caser冠军方案Caser-paddleWSDM 2018
Field-Embedded Factorization Machines for Click-through rate prediction[论文复现赛第4期][DeepFEFM冠军方案]DeepFEFM-paddle
Deep Learning Recommendation Model for Personalization and Recommendation Systems论文复现赛: 基于PaddleRec 24小时快速复现经典 CTR 预估算法
BERT4Rec:Sequential Recommendation with Bidirectional Encoder Representations from Transformer基于PaddlePaddle的BERT4Rec复现
Neural Collaborative Reasoning协同推理:基于飞桨复现NCR模型

🚀 人脸识别

论文精品项目顶会
ArcFace: Additive Angular Margin Loss for Deep Face Recognition人脸识别:基于Paddle框架的ArcFace论文复现CVPR 2019
RetinaFace: Single-stage Dense Face Localisation in the WildRetinaFace-Final
Real-time Convolutional Neural Networks for Emotion and Gender Classification情感和性别分类:基于飞桨复现SimpleCNN和MiniXception

🚀 小样本学习

论文精品项目顶会
Prototypical Networks for Few-shot Learning论文复现:ProtoNet-小样本分类领域经典之作NeurIPS 2017
Model-Agnostic Meta-Learning for Fast Adaptation of Deep NetworksMAML元学习算法,小样本学习,多任务学习ICML 2017

🚀 自监督学习

论文精品项目顶会
TabNet: Attentive Interpretable Tabular Learning表格学习:基于飞桨复现TabNet网络

🚀 姿态估计

论文精品项目顶会
Learning to See in the Dark图像增强:Learning to See in the DarkCVPR 2018
Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields第四期论文复现赛第一名代码方案-人体姿态估计(RMPose_PAFs)CVPR 2017
Stacked Hourglass Networks for Human Pose Estimation人体姿态估计:基于飞桨复现Hourglass

🚀 对抗攻击

论文精品项目顶会
Towards Deep Learning Models Resistant to Adversarial Attacks论文复现赛:对抗攻击ICLR 2018

🚀 OCR

论文精品项目顶会
Shape Robust Text Detection with Progressive Scale Expansion NetworkPseNet第三期论文复现赛第一名代码CVPR 2019
Detecting Text in Natural Image with Connectionist Text Proposal Network第四期复现挑战赛之文本检测算法ctpn复现ICLR 2018

🚀 贝叶斯推理

论文精品项目顶会
Weight Uncertainty in Neural Networks论文复现赛:贝叶斯神经网络
Logo

学大模型,用大模型上飞桨星河社区!每天8点V100G算力免费领!免费领取ERNIE 4.0 100w Token >>>

更多推荐