We introduce a guide to help deep learning practitioners understand and manipulate convolutional neural network architectures. The guide clarifies the relationship between various properties (input shape, kernel shape, zero padding, strides and output shape) of convolutional, pooling and transposed convolutional layers, as well as the relationship between convolutional and transposed convolutional layers. Relationships are derived for various cases, and are illustrated in order to make them intuitive.
A guide to convolution arithmetic for deep learning
A guide aids deep learning practitioners in understanding and manipulating convolutional neural network architectures, detailing relationships between layer properties like input and output shapes, zero padding, strides, and convolution and transposed convolution.
- Year
- 2016
- Venue
- arXiv 2016
- Authors
- 2
- Hosting
- Abstract onlyARXIV-DEFAULT
Cite
Notes
Only stored in your browser.
Attribution
- Abstract & full text
- arxiv.org/abs/1603.07285v2ARXIV-DEFAULT
- TL;DR
- Semantic Scholar