Statements (53)
Predicate | Object |
---|---|
gptkbp:instance_of |
gptkb:neural_networks
|
gptkbp:application |
Image Classification
|
gptkbp:architecture |
gptkb:Deep_Learning
|
gptkbp:based_on |
Inception Module
|
gptkbp:developed_by |
gptkb:François_Chollet
|
https://www.w3.org/2000/01/rdf-schema#label |
Xception
|
gptkbp:image_processing |
gptkb:stage_adaptation
|
gptkbp:input_output |
299x299
|
gptkbp:is_a_framework_for |
gptkb:Keras
|
gptkbp:language |
gptkb:Python
|
gptkbp:license |
MIT License
|
gptkbp:notable_feature |
Scalable architecture
Widely used in research Used in various competitions Efficient computation Flexible input size Pre-trained models available Compatible with Tensor Flow Can be deployed on mobile devices Can be used for anomaly detection Can be used for facial recognition Can be used for game AI Can be used for generative tasks Can be used for medical image analysis Can be used for natural language processing Can be used for object detection Can be used for recommendation systems Can be used for reinforcement learning Can be used for robotics Can be used for segmentation tasks Can be used for speech recognition Can be used for style transfer Can be used for time series forecasting Can be used for video analysis Good for fine-tuning High accuracy on large datasets Robust to overfitting Supports mixed precision training Supports model quantization Supports multi-GPU training |
gptkbp:orbital_period |
22,910,480
|
gptkbp:performance |
Top-1 Accuracy
Top-5 Accuracy |
gptkbp:predecessor |
gptkb:Inception-v3
|
gptkbp:provides_information_on |
gptkb:Image_Net
|
gptkbp:release_year |
gptkb:2017
|
gptkbp:successor |
gptkb:Efficient_Net
|
gptkbp:training |
Varies by hardware
|
gptkbp:uses |
Depthwise Separable Convolutions
|
gptkbp:uses_batch_normalization |
gptkb:True
|
gptkbp:uses_dropout |
gptkb:True
|
gptkbp:bfsParent |
gptkb:tf.keras.applications
|
gptkbp:bfsLayer |
5
|