|
gptkbp:instance_of
|
gptkb:neural_networks
|
|
gptkbp:application
|
gptkb:Computer_Vision
|
|
gptkbp:architecture
|
Image Classification
|
|
gptkbp:base_model
|
gptkb:Efficient_Net
|
|
gptkbp:collaborations
|
gptkb:Yes
|
|
gptkbp:community_support
|
Active
|
|
gptkbp:contribution
|
gptkb:Yes
|
|
gptkbp:developed_by
|
gptkb:Google_AI
|
|
gptkbp:drops
|
0.2
|
|
gptkbp:evaluates
|
Accuracy
Cross-Validation
|
|
gptkbp:feature_extractor
|
gptkb:Yes
|
|
gptkbp:has_publications
|
Efficient Net: Rethinking Model Scaling for Convolutional Neural Networks
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Efficient Net B1
|
|
gptkbp:hyperparameters
|
Tunable
|
|
gptkbp:image_classification_benchmark
|
Top-1 Accuracy on Image Net
|
|
gptkbp:image_input
|
RGB Images
|
|
gptkbp:impact
|
gptkb:High
|
|
gptkbp:inference_speed
|
Fast
|
|
gptkbp:initiated_by
|
Swish
|
|
gptkbp:input_output
|
Tensor
240x240
|
|
gptkbp:input_preprocessing
|
Normalization and Resizing
|
|
gptkbp:is_a_framework_for
|
gptkb:Tensor_Flow
|
|
gptkbp:is_optimized_for
|
gptkb:Adam
|
|
gptkbp:is_taught_in
|
0.001
|
|
gptkbp:layer_type
|
Depthwise Separable Convolution
|
|
gptkbp:license
|
Apache License 2.0
|
|
gptkbp:migration
|
Supported
|
|
gptkbp:model
|
gptkb:Efficient_Net
< 10 MB
B0, B2, B3, B4, B5, B6, B7
|
|
gptkbp:model_compression
|
Supported
|
|
gptkbp:model_deployment
|
Cloud and Edge
|
|
gptkbp:model_documentation
|
Available
|
|
gptkbp:model_download_link
|
gptkb:Tensor_Flow_Model_Garden
|
|
gptkbp:model_interpretability
|
Grad-CAM
|
|
gptkbp:model_updates
|
gptkb:Regular
|
|
gptkbp:model_visualization
|
gptkb:board_game
|
|
gptkbp:normalization
|
gptkb:Batch_Normalization
|
|
gptkbp:orbital_period
|
7.8 million
|
|
gptkbp:output_classes
|
1000
|
|
gptkbp:output_layer
|
Softmax
|
|
gptkbp:performance
|
FLOPs
|
|
gptkbp:predecessor
|
gptkb:Efficient_Net_B0
|
|
gptkbp:provides_information_on
|
gptkb:Image_Net
gptkb:Yes
|
|
gptkbp:related_model
|
Mobile Net, Res Net
|
|
gptkbp:release_year
|
gptkb:2019
|
|
gptkbp:scaling_method
|
Compound Scaling
|
|
gptkbp:sister_channel
|
gptkb:3
|
|
gptkbp:successor
|
gptkb:Efficient_Net_B2
|
|
gptkbp:top1_accuracy
|
79.6%
|
|
gptkbp:top5_accuracy
|
94.5%
|
|
gptkbp:training
|
gptkb:stage_adaptation
< 1 day
|
|
gptkbp:training_programs
|
gptkb:Keras
|
|
gptkbp:tuning
|
gptkb:Yes
|
|
gptkbp:use_case
|
Image Recognition
|
|
gptkbp:bfsParent
|
gptkb:Efficient_Net
|
|
gptkbp:bfsLayer
|
5
|