|
gptkbp:instanceOf
|
gptkb:convolutional_neural_network
|
|
gptkbp:activatedBy
|
gptkb:ReLU
|
|
gptkbp:basedOn
|
DenseNet architecture
|
|
gptkbp:developedBy
|
gptkb:Laurens_van_der_Maaten
gptkb:Gao_Huang
gptkb:Kilian_Q._Weinberger
gptkb:Zhuang_Liu
|
|
gptkbp:digestSize
|
224x224 pixels
|
|
gptkbp:frameworkAvailable
|
gptkb:TensorFlow
gptkb:Keras
gptkb:MXNet
gptkb:PyTorch
|
|
gptkbp:hasBlockType
|
Dense Block
Transition Layer
|
|
gptkbp:hasParameterCount
|
~8 million
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
DenseNet-121
|
|
gptkbp:level
|
121
|
|
gptkbp:normalization
|
Batch Normalization
|
|
gptkbp:notableFeature
|
feature reuse
dense connectivity
reduced vanishing-gradient problem
|
|
gptkbp:notablePublication
|
gptkb:Densely_Connected_Convolutional_Networks
|
|
gptkbp:outputClasses
|
1000 (ImageNet)
|
|
gptkbp:pooling
|
Average Pooling
|
|
gptkbp:publicationYear
|
2017
|
|
gptkbp:publishedIn
|
gptkb:CVPR_2017
|
|
gptkbp:trainer
|
gptkb:ImageNet
|
|
gptkbp:type
|
gptkb:feedforward_neural_network
|
|
gptkbp:usedFor
|
feature extraction
image classification
object detection
|
|
gptkbp:bfsParent
|
gptkb:DenseNet
|
|
gptkbp:bfsLayer
|
6
|