HyperAI超神经

首页资讯最新论文教程数据集百科 SOTA LLM 模型天梯 GPU 天梯顶会

中文

HyperAI超神经

Image Classification

图像分类是计算机视觉中的基本任务，旨在对整幅图像进行理解和归类，赋予其特定标签。该任务通常针对单个对象的图像，通过深度学习等技术实现高精度分类，具有广泛的应用价值，如内容识别、场景理解等。当分类达到实例级时，与图像检索相关联，后者还涉及在大型数据库中查找相似图像。

AP-GeM (ResNet-101)

µ2Net+ (ViT-L/16)

Certificate Verification

DINOv2 (ViT-g/14, frozen model, linear eval)

CIFAR-10 (40 Labels, ImageNet-100 Unlabeled)

CIFAR-10, 40% Symmetric Noise

CIFAR-10, 60% Symmetric Noise

CIFAR-10 Image Classification

CIFAR-10 (with noisy labels)

CIFAR-100, 40% Symmetric Noise

CIFAR-100, 60% Symmetric Noise

CIFAR-100 (alpha=0, 20 clients per round)

WRN-28-2 + UDA+AutoDropout

VIT-L/16 (Spinal FC, Background)

SEER (RegNet10B)

SEER (RegNet10B)

Clothing1M (using clean data)

ColonINST-v1 (Seen)

ColonINST-v1 (Unseen)

Colored-MNIST(with spurious correlation)

Entropy-based Logic Explained Network

ViT-Large/16 (384)

ViT-Large/16 (384)

Linear FT(ViT-L/14)

EMNIST-Balanced

WaveMixLite-128/7

µ2Net (ViT-L/16)

VGG-5(Spinal FC)

µ2Net+ (ViT-L/16)

Continued fraction of straight lines

TransBoost-ResNet50

EnGraf-Net101 (G=4, H=1)

Flowers (Tensorflow)

CNN+ Wilson-Cowan model RNN

FMD (materials)

TransBoost-ResNet50

LRA-diffusion (CLIP ViT)

Fracture/Normal Shoulder Bone X-ray Images on MURA

Our Ensemble Learning-2

Galaxy10 DECals

Fuzzy Distance Ensemble

Id Pattern Dataset

GAC-SNN MS-ResNet-34

ResNet-50 + UDA+AutoDropout

SparseSwin with L2

ImageNet-100 (Class-IL, 5T)

WRN (N=28, k=10)

WRN (N=36, k=5)

EfficientNet-L2-Ns

SqueezeNet + Simple Bypass

ViT-H @224 (DeiT III, 21k)

ImageNet-Sketch

µ2Net+ (ViT-L/16)

Model soups (ViT-G/14)

Imbalanced CUB-200-2011

WaveMix-256/16 (level 2)

AIMv2-3B (448 res)

iNaturalist 2018

iNaturalist 2019

Hiera-H (448px)

Intel Image Classification

ISIC 2018+Atlas Dermatology

iWildCam2020-WILDS

V-MoE-H/14 (Every-2)

SEER (RegNet10B)

µ2Net (ViT-L/16)

RADAM (ConvNeXt-XL)

Kuzushiji-MNIST

Large Labelled Logo Dataset (L3D)

L3D_original_2level

Malaria Dataset

kEffNet-B0 V2 16ch

EfficientNet-B3

mini WebVision 1.0

Branching/Merging CNN + Homogeneous Vector Capsules

PDO-eConv (ours)

MNIST-rot-12k (DA)

PDO-eConv (ours)

NCT-CRC-HE-100K

New Plant Diseases Dataset

No Background RGB Arabic Alphabets Sign Language Dataset

Noisy MNIST (AWGN)

Noisy MNIST (Contrast)

Noisy MNIST (Motion)

ObjectNet (Bounding Box)

ObjectNet (ImageNet classes)

Diffusion Classifier (zero-shot)

ResNet-18 + Vision Eagle Attention

Oxford-IIIT Pet Dataset

TWIST (ResNet-50)

Oxford-IIIT Pets

CeiT-S (384 finetune resolution)

PASCAL VOC 2007

MAE (ViT-H, 448)

InternImage-H（CNN）

Places365-Standard

SWAG (ViT H/14)

kMobileNet V3 Large 16ch

ResNet-152 2x (RS training)

Deep regularization

Red MiniImageNet 20% label noise

Red MiniImageNet 40% label noise

Red MiniImageNet 60% label noise

Red MiniImageNet 80% label noise

RGB Arabic Alphabet Sign Language (AASL) dataset

Fuzzy rank-based fusion of CNN models using Gompertz function

Heinsen Routing

split CIFAR-100

Split Fashion M-NIST

Model with negotiation paradigm

Max Margin Contrastive

Stanford Online Products

µ2Net+ (ViT-L/16)

TransBoost-ResNet50

Wide-ResNet-28-10

Tiered ImageNet 5-way (5-shot)

EGNN+Transduction

Tiny ImageNet Classification

touchtech/fashion-images-gender-age

Training and validation dataset of capsule vision 2024 challenge.

BiomedCLIP+PubmedBERT

Visual Wake Words

VizWiz-Classification

ALIGN (50 hypers/task)

CurriculumNet (InceptionResNet-v2)