Xception: Deep Learning with Depthwise Separable Convolutions

We present an interpretation of Inception modules in convolutional neural networks as being an intermediate step in-between regular convolution and the depthwise separable convolution operation (a depthwise convolution followed by a pointwise convolution). In this light, a depthwise separable convolution can be understood as an Inception module with a maximally large number of towers... (read more)

PDF Abstract

Datasets


  Add Datasets introduced or used in this paper

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Image Classification ImageNet Xception Top 1 Accuracy 79% # 101
Top 5 Accuracy 94.5% # 61
Number of params 22.8M # 69

Methods used in the Paper


METHOD TYPE
Average Pooling
Pooling Operations
Inception Module
Image Model Blocks
Residual Connection
Skip Connections
Pointwise Convolution
Convolutions
Dropout
Regularization
Weight Decay
Regularization
Step Decay
Learning Rate Schedules
RMSProp
Stochastic Optimization
SGD with Momentum
Stochastic Optimization
1x1 Convolution
Convolutions
Softmax
Output Functions
Dense Connections
Feedforward Networks
Global Average Pooling
Pooling Operations
Max Pooling
Pooling Operations
Depthwise Separable Convolution
Convolutions
ReLU
Activation Functions
Xception
Convolutional Neural Networks
Depthwise Convolution
Convolutions
Convolution
Convolutions