Statements (22)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
television channel
|
| gptkbp:composedOf |
3 fully connected layers
5 convolutional layers |
| gptkbp:developedBy |
NVIDIA
|
| gptkbp:exhibits |
the effectiveness of large datasets
|
| gptkbp:has |
over 60 million parameters
|
| https://www.w3.org/2000/01/rdf-schema#label |
LexNet
|
| gptkbp:illustratedBy |
data augmentation techniques
|
| gptkbp:influenced |
subsequent deep learning models
|
| gptkbp:inspiredBy |
AlexNet_architecture
|
| gptkbp:introduced |
2012
|
| gptkbp:is |
a landmark in deep learning history
a_convolutional_neural_network_(CNN) |
| gptkbp:operates |
Caffe framework
|
| gptkbp:providesAccessTo |
top-5 error rate of 16.4% on ImageNet
|
| gptkbp:significantEvent |
image recognition tasks
|
| gptkbp:usedFor |
image classification
|
| gptkbp:uses |
GPU_for_training
|
| gptkbp:utilizes |
ReLU activation function
dropout for regularization |
| gptkbp:vehicleClass |
deeper neural networks
|
| gptkbp:winner |
ILSVRC_2012
|