Long Tail Learning On Places Lt

Metriken

Top-1 Accuracy

Ergebnisse

Leistungsergebnisse verschiedener Modelle zu diesem Benchmark

Modellname	Top-1 Accuracy	Paper Title	Repository
LFME + OLTR	36.2	Learning From Multiple Experts: Self-paced Knowledge Distillation for Long-tailed Classification
IEM	39.7	Inflated Episodic Memory With Region Self-Attention for Long-Tailed Visual Recognition	-
DisAlign	39.3	Distribution Alignment: A Unified Framework for Long-tail Visual Recognition
OPeN (ResNet-152)	40.5	Pure Noise to the Rescue of Insufficient Data: Improving Imbalanced Classification by Training on Random Noise Images
BALLAD(ViT-B-16)	49.5	A Simple Long-Tailed Recognition Baseline via Vision-Language Model
VL-LTR (ResNet-50)	48.0	VL-LTR: Learning Class-wise Visual-Linguistic Representation for Long-Tailed Visual Recognition
CBExperts	38.9	Long-Tailed Recognition Using Class-Balanced Experts
LIFT (ViT-L/14)	53.7	Long-Tail Learning with Foundation Model: Heavy Fine-Tuning Hurts
Domain Adaptation	30.8	Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition from a Domain Adaptation Perspective
BALMS	38.7	Balanced Meta-Softmax for Long-Tailed Visual Recognition
BALLAD(ResNet-101)	47.9	A Simple Long-Tailed Recognition Baseline via Vision-Language Model
APA (SE-ResNet-50)	42.0	Adaptive Parametric Activation	-
LIFT (ViT-B/16)	52.2	Long-Tail Learning with Foundation Model: Heavy Fine-Tuning Hurts
LDAM-DRS-RSG	39.3	RSG: A Simple but Effective Module for Learning Imbalanced Datasets
VL-LTR (ViT-B-16)	50.1	VL-LTR: Learning Class-wise Visual-Linguistic Representation for Long-Tailed Visual Recognition
BALLAD(ResNet-50×16)	49.3	A Simple Long-Tailed Recognition Baseline via Vision-Language Model
BALLAD(ResNet-50)	46.5	A Simple Long-Tailed Recognition Baseline via Vision-Language Model
GPaCo (ResNet-152)	41.7	Generalized Parametric Contrastive Learning
MAM (ViT-B/16)	51.4	Improving Image Recognition by Retrieving from Web-Scale Image-Text Data	-
RAC (ViT-B-16)	47.17	Retrieval Augmented Classification for Long-Tail Visual Recognition	-

0 of 29 row(s) selected.