Code sample for Hoffman lab application.
The code trains a U-net autoencoder in a supervised fashion. Both the input and output are part of the .npy
file, provided in this repository.
Pytorch 1.9.0
Numpy 1.21.2
matplotlib 3.4.3
Load an anaconda pytorch environment.
Run the code as python DLHoffmanLabSample.py
Lossfile_train.txt
SavedParameters.pth
And a pair of images showing contact maps and distance maps after every two iterations, these files are named as follows
Autoencoder_contact<epoch number>.png
and Autoencoder_distance<epoch number>.png