Published July 9, 2019 | Version v1

Data files for In situ training of feedforward and recurrent convolutional memristor networks

Description

MATLAB data files for the manuscript "In situ training of feedforward and recurrent convolutional memristor networks" published on Nature Machine Intelligence, 2019.

The MATLAB data file "exp_mnist.mat" consists of all experimental data on implementing the convolutional neural network with the 1-transistor 1-memristor array that is used for plotting the Figure 1 and 2 of the manuscript.

The MATLAB data file "exp_mnistsequence.mat" consists of all experimental data on implementing the convolutional long short-term memory network on the 1-transistor 1-memristor array that is used for plotting the Figure 3 and 4 of the manuscript.

The code that generated these data files are provided by the link within the manuscript. Alternatively, the code can be accessed via https://github.com/zhongruiwang/memristorCNN.

Files

Files (1.6 GB)

Name Size
md5:ccf21779b25d739c56644ab2605fcc00
1.3 GB Download
md5:0088926864e1f29cf020b419cf119dfb
303.6 MB Download

Additional details

Related works

References

  • Wang, Zhongrui et al. (2019) In situ training of feed-forward and recurrent convolutional memristor networks, Nat. Mach. Intell.