Scribd
Upload
187 views14 pages

ZFNET Architecture

ZFNet is an 8-layer convolutional neural network architecture that improved upon AlexNet with 5 convolutional layers and 3 fully-connected layers. It takes in 224x224x3 pixel images as input and applies convolution and max pooling layers with various kernel sizes and strides before connecting to fully-connected layers with 4096, 4096, and 1000 neurons to produce predictions.

Uploaded by

Mehak Smagh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
187 views14 pages

ZFNET Architecture

ZFNet is an 8-layer convolutional neural network architecture that improved upon AlexNet with 5 convolutional layers and 3 fully-connected layers. It takes in 224x224x3 pixel images as input and applies convolution and max pooling layers with various kernel sizes and strides before connecting to fully-connected layers with 4096, 4096, and 1000 neurons to produce predictions.

Uploaded by

Mehak Smagh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

ZFNet(Matthew Zeiler and Rob

Fergus) architecture
• Significant improvement over Alexnet
architecture
• 8 layer convnet model
– 5 convolution layer & 3 FC layers
ZFNetarchitecture
• Input layer
– 224×224×3
• CL1 (96 kernels, Filter size :7×7, S=2)
Max pool (ML1)

3x 3 stride=2
CL2
• 5x5, s=2
• 256 kernels
ML2
• 3x3,s=2
CL3
• 3x3,pad=1,
• 384 kernels
CL4
• 3x3,pad=1,
• 384 kernels
CL5
• 3x3,pad=1,
• 256 kernels
ML3
• 3x3,s=2,
FC(1)
4096 neurons
FC(2)
4096 neurons
Output layer –softmax
FC(3)
1000 neurons
ZF Net architecture

https://arxiv.org/pdf/1311.2901.pdf
References
[1] https://arxiv.org/pdf/1311.2901.pdf

You might also like