TResNet: High Performance GPU-Dedicated Architecture

30 Mar 2020Tal RidnikHussam LawenAsaf NoyItamar FriedmanEmanuel Ben BaruchGilad Sharir

Many deep learning models, developed in recent years, reach higher ImageNet accuracy than ResNet50, with fewer or comparable FLOPS count. While FLOPs are often seen as a proxy for network efficiency, when measuring actual GPU training and inference throughput, vanilla ResNet50 is usually significantly faster than its recent competitors, offering better throughput-accuracy trade-off... (read more)

PDF Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
RESULT LEADERBOARD
Image Classification CIFAR-10 TResNet-XL Percentage correct 99 # 1
Image Classification CIFAR-100 TResNet-XL Percentage correct 91.5 # 4
Image Classification Flowers-102 TResNet-L Accuracy 99.1% # 3
Image Classification ImageNet TResNet-XL Top 1 Accuracy 84.3% # 22
Number of params 77M # 18
Multi-Label Classification MS-COCO TResNet-L mAP 86.4 # 1
Fine-Grained Image Classification Oxford 102 Flowers TResNet-L Accuracy 99.1% # 3
Fine-Grained Image Classification Stanford Cars TResNet-L Accuracy 96.0% # 2

Methods used in the Paper