gptkbp:instanceOf
|
deep neural network architecture
|
gptkbp:activatedBy
|
gptkb:ReLU
|
gptkbp:architecture
|
gptkb:convolutional_neural_network
|
gptkbp:benchmarkDataset
|
gptkb:COCO
gptkb:ImageNet
|
gptkbp:category
|
gptkb:artificial_intelligence
computer vision
deep learning
|
gptkbp:citation
|
over 100,000
|
gptkbp:developedBy
|
gptkb:Shaoqing_Ren
gptkb:Xiangyu_Zhang
gptkb:Jian_Sun
gptkb:Kaiming_He
|
gptkbp:digestSize
|
224x224 pixels
|
gptkbp:fullName
|
gptkb:Residual_Network
|
https://www.w3.org/2000/01/rdf-schema#label
|
ResNet
|
gptkbp:implementedIn
|
gptkb:TensorFlow
gptkb:Keras
gptkb:PyTorch
|
gptkbp:influenced
|
gptkb:DenseNet
gptkb:ResNeXt
gptkb:EfficientNet
|
gptkbp:introducedIn
|
gptkb:CVPR_2016
2015
|
gptkbp:level
|
101
152
18
34
50
|
gptkbp:license
|
gptkb:MIT_License
|
gptkbp:normalization
|
Batch Normalization
|
gptkbp:notableAchievement
|
first place in ILSVRC 2015 classification
first place in ILSVRC 2015 detection
|
gptkbp:notableFeature
|
alleviates vanishing gradient problem
identity shortcut connection
|
gptkbp:notableFor
|
skip connections
residual learning
|
gptkbp:notablePublication
|
gptkb:Deep_Residual_Learning_for_Image_Recognition
https://arxiv.org/abs/1512.03385
|
gptkbp:openSource
|
yes
|
gptkbp:parameter
|
25.6 million (ResNet-50)
44.5 million (ResNet-101)
60.2 million (ResNet-152)
|
gptkbp:relatedTo
|
gptkb:GoogLeNet
gptkb:VGGNet
gptkb:AlexNet
|
gptkbp:usedFor
|
image classification
object detection
semantic segmentation
|
gptkbp:bfsParent
|
gptkb:convolutional_neural_network
gptkb:ImageNet_competition
|
gptkbp:bfsLayer
|
5
|