Note
This means that the only configuration that is needed is setting the folders where pySkNet and SkyNet will read and write.
First clone the pySkyNet repository:
$ git clone https://github.com/cbonnett/SkyNet_wrapper.git
Then create a system variable SKYNETPATH to an exciting folder where you want to store SkyNet training, validation, predictions, configuration and prediction files.
Bash shell:
$ export SKYNETPATH=/path/where/you/want/to/store/skynet/data/
c-shell:
$ setenv SKYNETPATH /path/where/you/want/to/store/skynet/data/
Once the $SKYNETPATH is set, in the pySkyNet repo you run:
$ python src/install.py
This folder will contain the training and validation files.
This folder contains the configuration files used by SkyNet
This folder contains the learned weight files. This folder will also contain the predictions of the training and validation samples
In this folder all the predictions are printed.
Note
All these folders can be set when instantiating SkyNetRegressor or SkyNetClassifier class:
sn_reg = SkyNetRegressor(id=’identification’,input_root=’/my/folder/inputs’)