Statements (59)
| Predicate | Object |
|---|---|
| gptkbp:instance_of |
gptkb:television_channel
|
| gptkbp:bfsLayer |
4
|
| gptkbp:bfsParent |
gptkb:Deep_Lab
|
| gptkbp:based_on |
Inception model
|
| gptkbp:consists_of |
14 convolutional layers
|
| gptkbp:developed_by |
gptkb:François_Chollet
|
| gptkbp:has |
Pre-trained models available
|
| gptkbp:has_achievements |
State-of-the-art performance on Image Net
|
| https://www.w3.org/2000/01/rdf-schema#label |
Xception architecture
|
| gptkbp:is_compared_to |
gptkb:Res_Net
gptkb:Inception-v3 VGG Net |
| gptkbp:is_evaluated_by |
gptkb:LFW
gptkb:CIFAR-10 gptkb:SVHN gptkb:Stanford_Dogs_dataset gptkb:Celeb_A gptkb:Fashion_MNIST gptkb:CIFAR-100 Oxford Pets dataset |
| gptkbp:is_implemented_in |
gptkb:Graphics_Processing_Unit
gptkb:Keras |
| gptkbp:is_known_for |
Scalability
High accuracy Low latency High throughput Batch normalization Robustness to overfitting Dropout regularization Residual connections Efficiency in parameter usage Layer-wise learning rate adjustment |
| gptkbp:is_optimized_for |
GPU acceleration
TPU acceleration |
| gptkbp:is_part_of |
gptkb:Keras_Applications
gptkb:Tensor_Flow_Model_Garden |
| gptkbp:is_related_to |
Transfer learning
|
| gptkbp:is_supported_by |
gptkb:document
Research papers Community contributions Online tutorials |
| gptkbp:is_used_in |
Image segmentation
Feature extraction Data augmentation Hyperparameter tuning Fine-tuning Facial recognition Video analysis Adversarial training Object detection Generative models Style transfer Medical image analysis Image classification tasks |
| gptkbp:release_year |
gptkb:2017
|
| gptkbp:suitable_for |
Mobile applications
Real-time image processing |
| gptkbp:training |
gptkb:Image_Net_dataset
|
| gptkbp:uses |
Depthwise Separable Convolutions
|