-
Notifications
You must be signed in to change notification settings - Fork 5.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
【飞桨论文复现挑战赛(第五期)】榜单 #37401
Comments
【队名】:自动AC机 【序号】:84 【论文】:Convolutional Sequence to Sequence Learning 【状态】:报名 【repo链接】:https://github.com/robinbg/Paddle-ConvSeq2Seq |
【队名】:自动AC机 【序号】:3 【论文】:MnasNet: Platform-Aware Neural Architecture Search for Mobile 【状态】:报名 |
【队名】:自动AC机 【序号】:1 【论文】:Masked Autoencoders Are Scalable Vision Learners 【状态】:报名 |
【队名】:dudu 【序号】:27 【论文】:BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation 【状态】:报名 |
【队名】:txyugood 【序号】:31 【论文】:Learning to Adapt Structured Output Space for Semantic Segmentation 【状态】:报名 |
【队名】:秃样ToSample 【序号】:45 【论文】:RobustScanner: Dynamically Enhancing Positional Clues for Robust Text Recognition 【状态】:报名 【repo链接】:https://github.com/byhteng/RobustScanner-Paddle |
【队名】:秃样ToSample 【序号】:46 【论文】:From Two to One: A New Scene Text Recognizer with Visual Language Modeling Network 【状态】:报名 【repo链接】:https://github.com/byhteng/VisionLAN-Paddle |
【队名】:PaddleViT 【序号】:6 【论文】:CycleMLP: A MLP-like Architecture for Dense Prediction 【状态】:报名&提交 【repo链接】:https://github.com/BR-IDL/PaddleViT/tree/develop/image_classification/CycleMLP |
【队名】:若若若月寒 【序号】:92 【论文】:Universal Language Model Fine-tuning for Text Classification 【状态】:报名 【repo链接】:https://github.com/akari0216/Paddle-awdlstm |
【队名】:秃样ToSample 【序号】:49 【论文】: Attention on attention for image captioning 【状态】:报名 |
【队名】:异常数据 【序号】:85 【论文】:DeBERTa: Decoding-enhanced BERT with Disentangled Attention 【状态】:报名 |
【队名】:simuler 【序号】:104 【论文】: FNet: Mixing Tokens with Fourier Transforms 【状态】:报名 |
【队名】:UnseenMe 【序号】:3 【论文】:MnasNet: Platform-Aware Neural Architecture Search for Mobile 【状态】:报名 |
【队名】:白鸟非鸟 【序号】:17 【论文】:Fast RCNN 【状态】:报名 【repo链接】:https://github.com/sunlizhuang/FastRCNN_PaddlePaddle |
【队名】:白鸟非鸟 【序号】:20 【论文】:You Only Look One-level Feature 【状态】:报名 【repo链接】:https://github.com/sunlizhuang/YOLOF_PaddlePaddle |
【队名】:辣椒炒肉 【序号】:75 【论文】:NRPA: Neural Recommendation with Personalized Attention 【状态】:报名 |
【队名】:辣椒炒肉 【序号】:78 【论文】:Training Deep AutoEncoders for Collaborative Filtering 【状态】:报名 【repo链接】:https://github.com/PlumTao/Paddle-DeepRecommender |
【队名】:pikachu_wcg 【序号】:52 【论文】:Uniter: Learning universal image-text representations 【状态】:报名 |
【队名】:jvzhao 【序号】:54 【论文】:Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet? 【状态】:报名 |
【队名】:arbiter 【序号】:74 【论文】:TiSASRec: Time Interval Aware Self-Attention for Sequential Recommendation 【状态】:报名 |
【队名】:lsh冲冲冲 【序号】:34 【论文】:Category-Level Adversarial Adaptation for Semantic Segmentation using Purified Features 【状态】:报名 【repo链接】:repo链接 |
【队名】:lsh冲冲冲 【序号】:54 【论文】:Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet? 【状态】:报名 【repo链接】:repo链接 |
【队名】:arbiter 【序号】:74 【论文】:TiSASRec: Time Interval Aware Self-Attention for Sequential Recommendation 【状态】:提交 |
【队名】:有空多学习 【序号】:10 【论文】:Recurrent Models of Visual Attention 【状态】:报名 【repo链接】:https://github.com/cjh3020889729/The---Paddle---implement--RAM |
【队名】:Jordan2020 【序号】:1 【论文】:Masked Autoencoders Are Scalable Vision Learners 【状态】:报名 |
【队名】:Adream 【序号】:43 【论文】:Primitive Representation Learning for Scene Text Recognition 【状态】:提交 【repo链接】:https://github.com/OneYearIsEnough/PaddleOCR/tree/pren_juanbudong |
【队名】:异常数据 【序号】:85 【论文】:DeBERTa: Decoding-enhanced BERT with Disentangled Attention 【状态】:提交 |
【队名】:灯塔镇 【序号】:95 【论文】:Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism 【状态】:报名&提交 【repo链接】:https://github.com/Beacontownfc/MegatronBert_paddle |
【队名】:xushouyan88的团队 【序号】:5 【论文】:Cream of the Crop: Distilling Prioritized Paths For One-Shot Neural Architecture Search 【状态】:报名 |
【队名】:paddle_as_torch 【序号】:89 【论文】:ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training 【状态】:报名&提交 【repo链接】:https://github.com/d294270681/ProphetNet-paddle |
【队名】:Ming Sung 【序号】:62 【论文】:Deeper Depth Prediction with Fully Convolutional Residual Networks 【状态】:提交 |
【队名】:重在参与 【序号】:1 【论文】:Masked Autoencoders Are Scalable Vision Learners 【状态】:提交 |
【队名】:丰收厅的斯莱特林 【序号】:105 【论文】:A Nyström-Based Algorithm for Approximating Self-Attention 【状态】:提交 【repo链接】:https://github.com/JeremyZhao1998/Nystromformer-Paddle |
【队名】:角灰大帝硬train一发 【序号】:84 【论文】:Convolutional Sequence to Sequence Learning 【状态】:提交 【repo链接】:https://github.com/MiuGod0126/ConvS2S_Paddle |
【队名】:湖工起飞 【序号】:206 【论文】:ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation 【状态】:提交 【repo链接】:PaddlePaddle/PaddleSeg#1769 |
【队名】:dudu 【序号】:6 【论文】:CycleMLP: A MLP-like Architecture for Dense Prediction 【状态】:报名&提交 |
【队名】:LYyyds 【序号】:76 【论文】:Detecting Beneficial Feature Interactions for Recommender Systems 【状态】:提交 |
【队名】:LYyyds 【序号】:82 【论文】:CAN: Revisiting Feature Co-Action for Click-Through Rate Prediction 【状态】:提交 |
【队名】:这道题我不会做呀 【序号】:7 【论文】:HRFormer: High-Resolution Transformer for Dense Prediction 【状态】:提交 【repo链接】:https://github.com/huilin16/HRFormer_paddle |
【队名】:fuqianya 【序号】:50 【论文】:VinVL: Revisiting Visual Representations in Vision-Language Models 【状态】:提交 |
【队名】:图克塔米舍夫 【序号】:102 【论文】: ByT5: Towards a token-free future with pre-trained byte-to-byte models 【状态】:提交 |
【队名】:linqin2011 【序号】:2 【论文】: Swin Transformer V2: Scaling Up Capacity and Resolution 【状态】:提交 【repo链接】:https://github.com/Qdriving/Swin-TransformerV2-in-PP.git |
1 similar comment
【队名】:linqin2011 【序号】:2 【论文】: Swin Transformer V2: Scaling Up Capacity and Resolution 【状态】:提交 【repo链接】:https://github.com/Qdriving/Swin-TransformerV2-in-PP.git |
【队名】:fuqianya 【序号】:52 【论文】:Uniter: Learning universal image-text representations 【状态】:报名&提交 |
【队名】:角灰大帝硬train一发 【序号】:255 【论文】:Recurrent Models of Visual Attention 【状态】:提交 |
【队名】:角灰大帝硬train一发 【序号】:239 【论文】:Mixer: An all-MLP Architecture for Vision 【repo链接】:https://github.com/MiuGod0126/Mlp-Mixer-Paddle |
【队名】:角灰大帝硬train一发 【序号】:90 |
【队名】:小吴今天运动了吗 【序号】:56 【论文】:DeeperCut: A Deeper, Stronger, and Faster Multi-Person Pose Estimation Model 【状态】:提交 【复现指标】:PCK=87.8 【repo链接】:https://github.com/Wuxiao85/Deeper-Pose-Estimation-Model |
【队名】:若若若月寒 【序号】:92 【论文】:Universal Language Model Fine-tuning for Text Classification 【状态】:提交 【repo链接】:https://github.com/akari0216/Paddle-awdlstm |
【队名】:为往圣继绝学 【序号】:4 【论文】:XCiT: Cross-Covariance Image Transformers 【状态】:提交 【repo链接】:https://github.com/BrilliantYuKaimin/XCiT-PaddlePaddle |
hi,大家好,非常高兴的告诉大家,百度飞桨论文复现赛第五期已经开始了,本次论文复现赛共将有100篇的经典&前沿论文供大家复现,以及新增了工程落地赛,详细信息可以参考AI Studio,大家是否已经迫不及待了呢~
为了帮助大家更好的了解每篇论文的复现进度,本次复现赛将在本帖中汇总信息。如果你报名/提交结果,需要按格式在进行回复,我们每天会汇总所有的信息,并更新在下表中,你可以通过下面的表格中,了解到每一篇论文的相关信息。
回复的格式为:
【队名】:你队伍的名称
【序号】:你队伍想要复现的论文序号
【论文】:你队伍想要复现的论文名称
【状态】:报名/提交
【repo链接】:repo的链接
如:
【队名】:百度飞桨
【序号】:100
【论文】:Paper 1
【状态】:提交
【repo链接】:https://github.com/PaddlePaddle/Paddle
论文复现赛
Image Classification(图像分类)
1Masked Autoencoders Are Scalable Vision LearnersJordan2020(报名)(MAE-Paddle)
初学者(报名&提交)(PaddleMAE)
sosojust(报名)(Paddle_MAE)
幼儿园大班(报名)(MAE_PADDLE)
风车村(报名)(MaskedAutoencodersAreScalableVisionLearners)
AE86(报名)(repo的链接)
重在参与(报名&提交)(mae-paddle)
云天的智能小屋(报名)(Paddle_MAE)
小白复现队(报名)(Paddle-MAE)
SmirnovKol(报名)(Paddle-MAE)
单人旅途(报名)(MAE-Paddle)
潜龙勿用(报名)(swinv2)
掉了钱眼里去了(报名)(SwinTf_V2_Paddle)
CrashKing(报名)(Paddle-SwinTransformerV2)
linqin2011(报名&提交)(Swin-TransformerV2-in-PP.git)
SmirnovKol(报名)(PaddlePaddle-Swin-Transformer-V2)
3MnasNet: Platform-Aware Neural Architecture Search for MobileUnseenMe(报名)(mnasnet-paddle)
东方幻翎(报名)(Paddle_MnasNet)
beyond(报名)(Paddle-MNAS)
SmirnovKol(报名)(PaddlePaddle-MnasNet)
arbiter(报名&提交)(MnasNet-paddle)
4XCiT: Cross-Covariance Image Transformers角灰大帝硬train一发(报名)(XCiT_Paddle)
Duer(报名)(Paddle-XCiT)
SmirnovKol(报名)(PaddlePaddle-XCiT)
十一(报名)(main)
SmirnovKol(报名)(PaddlePaddle-Cream-of-the-Crop)
xushouyan88的团队(报名)(cream)
SmirnovKol(报名)(PaddlePaddle-CycleMLP)
十一(报名&提交)(CycleMLP-paddle)
dudu(报名&提交)(CycleMlp_paddle)
这道题我不会做呀(报名&提交)(HRFormer_paddle)
SmirnovKol(报名)(PaddlePaddle-HRFormer)
会飞的明湖烤鸭(报名)(PVTv2_paddle)
SmirnovKol(报名)(PaddlePaddle-PVTv2)
辉辉的战斗鸡(报名)(Paddle-PVT-V2)
dudu(报名&提交)(PVTV2_paddle)
SmirnovKol(报名)(Emerging-Properties-in-Self-Supervised-Vision-Transformers)
10Recurrent Models of Visual Attention马骏骁(报名)(Recurrent-Models-of-Visual-Attention)
角灰大帝硬train一发(报名&提交)(RAM-Paddle)
三根火柴(报名)(-RNN-)
SmirnovKol(报名&提交)(Recurrent-Models-of-Visual-Attention-using-paddlepaddle)
晓飞好队(报名)(recurrent-visual-attention-paddle)
jsdbzcm(报名)(Recurrent-Models-of-Visual-Attention)
QIQ(报名)(Paddle-DEIT)
SmirnovKol(报名&提交)(Going-deeper-with-Image-Transformers-using-PaddlePaddle)
PCALab-NNDM小组(报名)(Paddle_gdiT)
12Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet1004-1210(报名)(Tokens-to-Token-ViT-for-paddlle)
114战队(报名)(T2T-ViT_Paddle)
填空题全队(报名)(T2T_Vit_paddle)
PCALab-NNDM小组(报名&提交)(T2T_paddle)
噫吁嚱(报名)(CvT_Paddle)
牛奶奥利奥(报名)(Paddle-CvT)
我要发CVPR(报名)(CVT)
GalaxyGYP(报名)(ThesisRetrieval)
没事找事干(报名)(Paddle-CvT)
菜鸟来了(报名)(CVT)
SmirnovKol(报名&提交)(PaddlePaddle-CvT)
桂小太郎(报名)(CvT_lizn.git)
PCALab-NNDM小组(报名)(Paddle_cvt)
14Rethinking Spatial Dimensions of Vision Transformers三鲜面哈哈(报名)(paddle-pit)
ssrYYYY(报名)(pit)
MARS(报名)(PiT)
shiyi(报名)(repo的链接)
HBU_ICS_MNBX(报名)(PiT)
胖蘑菇蟹蟹(报名)(paddle_pit)
SmirnovKol(报名)(PaddlePaddle-PiT)
114战队(报名&提交)(paddle_pit)
Object Detection(目标检测)
15VoxelNet: End-to-End Learning for Point Cloud Based 3D Object DetectionSmirnovKol(报名)(PaddlePaddle-VoxelNet)
波波啥也不会(报名 &提交)(VoxelNet)
牛牛向前冲(报名)(Paddle_RetinaMask.git)
SmirnovKol(报名)(PaddlePaddle-RetinaMask)
PangAY(报名)(RetinaMask)
just4fun(报名)(Paddle_RetinaMask)
原魔玩家(报名)(paddle_retinamask)
起飞(报名)(RetinaMask-paddle)
17Fast R-CNN同路人(报名)(Fast-R-CNN)
优雅永不过时(报名&提交)(repo的链接)
双水冗替(报名)(FastRCNN_PaddlePaddle)
222333(报名)(faster-r-cnn-recurrent)
18Deep Predictive Coding Networks for Video Prediction and Unsupervised Learningarbiter(报名&提交)(prednet-paddle)
龍襲22的团队(报名)(paddle_DetectoRS)
Fitzzzzz(报名)(PaddleDetectoRS)
thestone(报名)(Paddle-DetectoRS)
SmirnovKol(报名)(PaddlePaddle-DetectoRS)
PangAY(报名)(DetectoRS)
just4fun(报名)(Paddle_DetectoRS)
20You Only Look One-level Feature上汽帆一ai小伙伴(报名)(igear-yolof-pdpd)
自信放光芒(报名)(Yolof-paddlepaddle)
114战队(报名)(YOLOF_Paddle)
shier(报名)(paddlepaddle-yolof)
牛牛向前冲(报名)(PaddlePaddle_YOLOF.git)
Copy is all you need(报名)(reproduce_yolo_with_paddle)
填空题全队(报名)(YOLOF-Paddle)
thestone(报名&提交)(Paddle-YOLOF)
21Gradient Harmonized Single-stage Detector芝麻开花(报名)(GHM_Detection_Paddle)
thestone(报名&提交)(Paddle-GHM)
牛牛向前冲(报名)(PaddlePaddle_GHSSD.git)
22YOLOX: Exceeding YOLO Series in 2021xihuanafeng(报名)(YOLOX-Paddle)
xihuanafeng(报名)(YOLOX-Paddle)
百度飞桨(报名)(paddle_yolox)
杨枝甘露(报名)(Paddle-YOLOX)
女装大佬咩酱(报名)(Paddle-YOLOX)
智科小队(报名)(YOLOX_paddle)
WhiteBlue(报名)(Paddle_YOLOX)
啥都不会团队(报名)(repo的链接)
DearPuff(报名)(paddle-yolox_repo)
牛牛向前冲(报名)(PaddlePaddle_YOLOX.git)
皮卡丘,12月飞行(报名)(YOLOX)
thestone(报名)(Paddle-YOLOX)
itInterpret(报名)(YOLOX_PP)
小白复现之路(报名)(YOLOX_Paddle_repo)
Kok(报名)(paddle)
HaoHao(报名&提交)(YOLOX-Paddle)
SmirnovKol(报名)(PaddlePaddle-YOLOX)
F1(报名)(pp_yolox)
就tm你叫人工智能(报名)(paddle-Instances_as_Queries)
喵呜(报名)(InstancesAsQueries_Paddle)
SmirnovKol(报名)(Instances-as-Queries)
just4fun(报名)(Paddle_QueryInst)
起飞(报名)(QueryInst-paddle)
Image Segmentation(图像分割)
beyond(报名)(paddle-vvs)
SmirnovKol(报名)(Video-to-Video-Synthesis)
24V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation不中CVPR不改名(报名)(Repo)
Fly(报名)(V-Net)
被褐怀玉的团队(报名)(Paddle_V-Net)
萬事勝意(报名)(Vnet-paddlepaddle)
被褐怀玉的团队(报名)(Paddle_V-Net)
萬事勝意(报名)(Vnet-paddlepaddle)
beyond(报名)(paddle-Vnet)
PangAY(报名)(V-Net)
SmirnovKol(报名)(V-Net-PaddlePaddle)
桂小太郎(报名)(Vnet_lizn)
25Context Encoding for Semantic Segmentation起飞(报名)(ENCNet-paddle)
波波啥也不会(报名)(EncNet-Paddle)
dudu(报名&提交)(ENCNet_paddle)
26Dynamic Multi-Scale Filters for Semantic Segmentationdudu(报名&提交)(DMNet_paddle)
27BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation初学者(报名)(PaddleBiSeNet)
flyingwheel(报名)(pp-BiSeNet)
鹅鹅鹅(报名)(BiSeNet-Paddle)
28Adversarial Learning for Semi-Supervised Semantic Segmentation波波啥也不会(报名&提交)(AdvSemiSeg-Paddle)
29Learning to Adapt Structured Output Space for Semantic Segmentation生活所迫(报名&提交)(AdaptSegNet-paddle)
jsdbzcm(报名)(AdaptSegNet)
30FastFCN: Rethinking Dilated Convolution in the Backbone for Semantic Segmentation今年还没秃(报名)(FastFCN_Paddle)
dudu(报名&提交)(FastFCN_paddle)
31ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentationsimuler(报名&提交)(ESPNet)
ABCD(报名)(paddle-DPT)
beyond(报名)(paddle-dpt)
Idolmaster(报名)(BTCV-Challenge)
Sousey冲鸭(报名)(Thesis-retrieval-Sousey)
wise(报名)(wise-UNETR)
beyond(报名)(paddle_UNETR)
小丑不丑(报名&提交)(Paddle-UNETR)
34Category-Level Adversarial Adaptation for Semantic Segmentation using Purified FeaturesImage Generation(图像生成)
雀氏蟀(报名)(Style-Based-GAN)
逗比小孩(报名)(Paddle_Image_Generation)
明光楼彭于晏(报名)(paddle_stylegan2_ada)
我们还行吧,那当然(报名)(paddle_stylegan2_ada)
峡谷见(报名)(GPEN)
桂小太郎(报名)(GPEN__lizn)
RandomFly(报名&提交2)(paddle-VQGAN)
yyds(报名)(StyleCLIP_Paddle)
GAN就完了(报名)(StyleCLIP-paddle.git)
40Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data滴滴滴(报名)(paddlefuxian)
勇敢土豆不怕困难!(报名&提交)(Real_ESRGAN_paddle)
AlexZou(报名)(RCAN_paddle)
只想摸鱼(报名&提交)(Paddle-RCAN)
他们让我报的,我不会啊(报名)(paddle-RCAN)
Key.L刘波(报名&提交)(RCAN-Paddle)
下雨不愁(报名)(Paddle-RCAN)
清水摸鱼(报名)(RCAN_Paddle)
云(报名)(Paddle-RCAN)
红糖糍粑(报名&提交)(PixelCNN)
桂小太郎(报名)(PixelCNN_lizn)
Scene Text
43Primitive Representation Learning for Scene Text Recognition6个6(报名)(paddle_PROCR)
Adream(报名&提交)(pren_juanbudong)
beyond(报名)(PP_RobustScanner)
Adream(报名)(rs_juanbudong)
IWIN-dgw(报名)(Vision-LAN-paddle)
Adream(报名)(vlan_juanbudong)
PeachML(报名)(VisionLAN.git)
beyond(报名)(pp_LGPMA)
48Fourier Contour Embedding for Arbitrary-Shaped Text Detectionzzm(报名&提交)(FCENet_Paddle)
Multi-Modal(多模态)
49Attention on attention for image captioningfuqianya(报名&提交)(AoANet-Paddle)
小丑不丑(报名&提交)(Paddle-AoA-Captioning)
Dandelight(报名)(AoAnet-PaddlePaddle)
50VinVL: Revisiting Visual Representations in Vision-Language Modelsfuqianya(报名&提交)(VinVL-Paddle)
51From Recognition to Cognition: Visual Commonsense Reasoning123123(报名)(Uniter-Paddle)
fuqianya(报名&提交)(UNITER-Paddle)
其他CV任务
53Drop an Octave: Reducing Spatial Redundancy in Convolutional Neural Networks with Octave ConvolutionSOLO(报名)(octconv-paddle)
Doctor Who(报名)(Paddle-OctConv)
arbiter(报名&提交)(OctConv-paddle)
54Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?lsh冲冲冲(报名)(repo链接)
生活所迫(报名&提交)(3D-ResNets-Paddle)
冲(报名)(3D_Resnet)
dudu(报名&提交)(3D-ResNets_paddle)
55Learning Spatiotemporal Features with 3D Convolutional Networkstxyugood(报名&提交)(PaddleC3D)
mmma-mua(报名)(DeeperCut-Paddle)
57Simple Baselines for Human Pose Estimation and Tracking波波啥也不会(报名&提交2)(HumanPoseBL-Paddle)
59Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks科研开小差(报名&提交)(MTCNN_paddle)
Ming Sung(报名&提交)(fcrn_paddle)
63Unsupervised Monocular Depth Estimation with Left-Right ConsistencySOLO(报名&提交)(monodepth-paddle)
64Digging Into Self-Supervised Monocular Depth EstimationSOLO(报名&提交)(monodepth2-paddle)
Ming Sung(报名)(paddle_monoDepth2)
两个祺(报名)(simsiam_paddle)
66FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Tracking67Finding Tiny Faceslsh冲冲冲(报名)(finding tiny faces)
HaoHao(报名&提交)(Tiny-Face-Paddle)
68MVFNet: Multi-View Fusion Network for Efficient Video Recognitiontxyugood(报名&提交)(PaddleMVF)
Recommendation Systems(推荐系统)
70DSelect-k: Differentiable Selection in the Mixture of Experts with Applications to Multi-Task LearningA3(报名&提交)(mtl-dselect-k)
71Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social RecommendationA3(报名&提交)(recall-mhcn)
72DCN V2: Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systems划划水(报名)(DCNv2-Paddle)
LYyyds(报名&提交)(DCN_V2_Paddle)
simuler(报名&提交)(DCN-V2)
74TiSASRec: Time Interval Aware Self-Attention for Sequential RecommendationPaddorch(报名&提交)(TiSASRec.paddle)
USMyyds(报名&提交)(Paddle-NRPA)
云图计划天下第一(报名)(Paddle-NRPA)
cHEn(报名)(Paddle-SIGN)
云图计划天下第一(报名)(Paddle-SIGN)
BamTeam(报名)(Paddle-SIGN)
77Efficient Non-Sampling Factorization Machines for Optimal Context-Aware Recommendation78Training Deep AutoEncoders for Collaborative FilteringKazeruJoy(报名)(DAE4CF)
HPU-三人成行(报名)(TiSASRec-paddle)
清华深圳SIGS(报名)(DeepRecommender_Paddle)
cHEn(报名&提交)(master)
皮卡丘,12月飞行(报名)(FGCNN_Studio)
大AI无疆(报名)(Paddle-FGCNN)
cHEn(报名&提交)(paddle-FGCNN)
80FLEN: Leveraging Field for Scalable CTR PredictioncHEn(报名)(paddle-FLEN)
81AutoFIS: Automatic Feature Interaction Selection in Factorization Models for Click-Through Rate Predictionarbiter(报名&提交)(autofis)
BamTeam(报名)(Paddle-CAN)
LYyyds(报名&提交)(CAN_Paddle)
lfy(报名)(DSIN)
划划水(报名)(Paddle-DSIN)
NLP(自然语言处理)
84Convolutional Sequence to Sequence Learning浪费时间(报名)(ConvSeq)
梨花春雪风舞的团队(报名)(CStSL)
灯塔镇(报名)(paddle_conv_seq2seq)
YBS(报名)(ConvS2S.git)
丰收厅的斯莱特林(报名)(CovSeq2Seq-Paddle)
角灰大帝硬train一发(报名&提交)(ConvS2S_Paddle)
异常数据(报名&提交)(DeBERTa)
还有家用电器(报名)(pd-DeBERTa)
布鲁克林篮网队(报名)(kevindurant)
开福区老实人(报名)(paddle-debertav2)
86Dense Passage Retrieval for Open-Domain Question Answering87I-BERT: Integer-only BERT Quantization金州勇士队(报名)(Paddle_Transformer)
浪费时间(报名)(Longformer)
丰收厅的斯莱特林(报名)(Longformer-Paddle)
89ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-trainingpaddle_as_torch(报名)(ProphetNet)
不知道起啥名就这样吧(报名)(paddle_ProphetNet.git)
paddle_as_torch(报名&提交)(ProphetNet-paddle)
角灰大帝硬train一发(报名&提交)(ConvS2S_Paddle)
角灰大帝硬train一发(报名)(FSMT_Paddle)
92Universal Language Model Fine-tuning for Text Classification丰收厅的斯莱特林(报名)(ULMFiT-Paddle)
不会怎么办(报名)(ULMFiT_paddle)
灯塔镇(报名)(adwlstm_paddle)
93Bag of Tricks for Efficient Text Classification94Multilingual Denoising Pre-training for Neural Machine Translation95Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism布鲁克林篮网队(报名)(kevindurant)
灯塔镇(报名&提交)(MegatronBert_paddle)
99LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention若若若月寒(报名)(Paddle-LUKE)
图克塔米舍夫(报名)(Paddle-LUKE)
灯塔镇(报名&提交)(paddle_luke_stable)
100RemBERT: Rethinking embedding coupling in pre-trained language modelsmask(报名)(Paddle-RemBERT)
若若若月寒(报名)(Paddle-RemBERT)
道不远人55(报名)(Paddle--RemBERT)
灯塔镇(报名&提交)(rembert_paddle)
arbiter(报名&提交1验收失败)(adalm-paddle)
102ByT5: Towards a token-free future with pre-trained byte-to-byte models图克塔米舍夫(报名&提交)(Paddle-ByT5)
104FNet: Mixing Tokens with Fourier Transforms神仙会(报名)(fnet_paddle)
手算神经网络(报名&提交)(Paddle-FNet)
105Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention丰收厅的斯莱特林(报名&提交)(Nystromformer-Paddle)
106Listen, Attend and Spell工程落地赛
项目地址:EfficientDet
项目地址:ScaledYOLOv4
203论文名称:Focal Loss for Dense Object Detection项目地址:RetinaNet
项目地址:YOLOv4
项目地址:htc
清水摸鱼(报名)(htc_paddle)
206论文名称:ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation项目地址:enet
湖工起飞(报名&提交)(PaddleSeg#1769)
Yan(报名)(Paddle_monodepth)
项目地址:hed
Yan(报名)(Paddle_monodepth2)
SmallWind(报名)(hed)
项目地址:UperNet
项目地址:deeplab
项目地址:BackgroundMattingV2
211论文名称:Panoptic Feature Pyramid Networks项目地址:panopticFPN
项目地址:CIC
项目地址:ProtoNet
项目地址:VA-Captioning
项目地址:RMPose_PAF
项目地址:BUTD-Captioning
项目地址:Pose
项目地址:SupContrast
项目地址:yolact2
项目地址:ContextPrior
项目地址:CLAN
项目地址:PartialConv
项目地址:T2I
224项目地址:SinGAN
项目地址:SPADE
项目地址:pSp
为往圣继绝学(报名)(pixel2style2pixel-PaddleGAN)
项目地址:SAM
项目地址:tabnet
项目地址:RetinaFace
项目地址:ctpn
项目地址:FaceClassification
项目地址:APCNet
项目地址:relational-networks
项目地址:CGNet
项目地址:ICNet
项目地址:DiscoGAN
项目地址:Old2Life
项目地址:PANet
项目地址:Mlp-Mixer
项目地址:blenderbot
吴先生(报名)(blenderbot_paddle)
241论文名称:ConvBERT: Improving BERT with Span-based Dynamic Convolution项目地址:convbert
项目地址:MobileBert_paddle
项目地址:t5
244论文名称:Field-Embedded Factorization Machines for Click-through rate prediction项目地址:deepfefm
245论文名称:Deep Learning Recommendation Model for Personalization and Recommendation Systems项目地址:DLRM
246项目地址:BERT4Rec
247论文名称:Fast R-CNN项目地址:repo的链接
248论文名称:Context Encoding for Semantic Segmentation项目地址:ENCNet_paddle
249论文名称:Dynamic Multi-Scale Filters for Semantic Segmentation项目地址:DMNet_paddle
250论文名称:BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation项目地址:BisNetV1_paddle
251论文名称:FastFCN: Rethinking Dilated Convolution in the Backbone for Semantic Segmentation项目地址:FastFCN_paddle
252论文名称:Adversarial Learning for Semi-Supervised Semantic Segmentation项目地址:AdvSemiSeg-Paddle
项目地址:AdaptSegNet-paddle
254论文名称:ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation项目地址:ESPNet
项目地址:RAM-Paddle
角灰大帝硬train一发(报名&提交)(RAM-Paddle)
256论文名称:Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning项目地址:prednet-paddle
项目地址:Paddle-GHM
项目地址:master
项目地址:T2T_paddle
项目地址:Real_ESRGAN_paddle
261论文名称:Fourier Contour Embedding for Arbitrary-Shaped Text Detection项目地址:FCENet_Paddle
项目地址:HumanPoseBL-Paddle
项目地址:MTCNN_paddle
项目地址:Tiny-Face-Paddle
项目地址:Paddle-YOLOF
项目地址:autofis
项目地址:paddle_r2c
项目地址:OctConv-paddle)
项目地址:paddle_pit
项目地址:monodepth-paddle
271论文名称:DSelect-k: Differentiable Selection in the Mixture of Experts with Applications to Multi-Task Learning项目地址:mtl-dselect-k
项目地址:YOLOX-Paddle
项目地址:VoxelNet
项目地址:Paddle-FNet
项目地址:recall-mhcn
项目地址:FLEN-Paddle
项目地址:DCN_V2_Paddle
项目地址:monodepth2-paddle
项目地址:mae-paddle
项目地址:paddle_PROCRpren_juanbudong
项目地址:MegatronBert_paddle
项目地址:Nystromformer-Paddle
项目地址:ProphetNet-paddle
项目地址:XCiT-PaddlePaddle
The text was updated successfully, but these errors were encountered: