Classification task for predicting landslides based on GIS maps using locally aligned convolutional neural networks. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License.
numpy
torch
sacred
tensorboard
matplotlib
sklearn
h5py
optim
: choice of the optimizerlr
: learning raten_epochs
: number of epochs to train the modelbs
: batch sizedecay
: L2 regularization parameter for the optimizerpatience
: number of epochs to wait before changing the learning rate in the schedulerpos_weight
: positive sample weight in the loss functionmodel
: model name to use for trainingn_workers
: number of workers to use for loading the data in data loaderregion
: the region of the datasetpix_res
: resolution of the pixels in the datasetstride
: stride of the CNNws
: window/kernel size of the CNNpad
: padding size in CNNfeature_num
: total number of featuresoversample
: boolean value indicating whether we want to oversample the data or notprune
: prunning size for the input imagesdist_num
: number of distance features (how far do we want to look)dist_feature
: boolean value indicating if we want to use distance features or notload_model
: path to the modeldata_path
: path to the dataindex_path
: path to the indices showing the partitioning of the datasave
: how often (how many epochs) we want to save the training models
Train the model using sacred specifying the arguments if you don't want to use the default values:
python main.py with 'train_params.optim="SGD"' ... 'loc_param.data_path="/tmp/Veneto_data.h5"' -m CNN
Validate the model:
python validate.py with 'params.load_model="/tmp/m1.pt"' ...
Tensorbaord Documentation for pytorch
This repository implements the paper "Predicting Landslides Using Locally Aligned Convolutional Neural Networks, A. Hajimoradlou, G. Roberti, and D. Poole". Please cite our paper, if you use the data or the code provided here in your work.
Links:
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License.