Properties (52)
Predicate | Object |
---|---|
gptkbp:instanceOf |
gptkb:Cloud_Computing_Service
|
gptkbp:appliesTo |
Depthwise separable convolutions
|
gptkbp:basedOn |
Neural Architecture Search
|
gptkbp:completed |
Higher accuracy with fewer parameters
State-of-the-art_accuracy_on_ImageNet |
gptkbp:composedOf |
EfficientNet-B0
EfficientNet-B1 EfficientNet-B2 EfficientNet-B3 EfficientNet-B4 EfficientNet-B5 EfficientNet-B6 EfficientNet-B7 |
gptkbp:developedBy |
gptkb:Google_AI
|
gptkbp:has |
Multiple variants
|
https://www.w3.org/2000/01/rdf-schema#label |
EfficientNet
|
gptkbp:improves |
Model efficiency
|
gptkbp:introduced |
2019
|
gptkbp:isAvailableIn |
gptkb:PyTorch
gptkb:Keras TensorFlow |
gptkbp:isChallengedBy |
Transfer learning
Image recognition Model compression Fine-tuning models Visual recognition tasks |
gptkbp:isEvaluatedBy |
gptkb:CIFAR-10
gptkb:Stanford_Dogs Open Images CIFAR-100 Fashion-MNIST Tiny ImageNet MS COCO VOC 2012 Food-101 Places365 Oxford_Pets |
gptkbp:isOptimizedFor |
Mobile devices
Cloud-based applications Edge devices |
gptkbp:isPartOf |
gptkb:ResNet
gptkb:MobileNet Inception DenseNet |
gptkbp:isSupportedBy |
CPUs
GPUs TPUs |
gptkbp:isUsedIn |
Image classification tasks
Object detection tasks Semantic segmentation tasks |
gptkbp:reduces |
Computational cost
|
gptkbp:uses |
Compound scaling
|