Branch Descriptions#
Year: 2023#
Modified AUM-Dataset_Latest (7/7/2023) : shighton/DeepHSI#
Added Dr. Kursun’s fixes from his repository (olcaykursun/DeepHyperX) to AUM-Dataset_Latest (shighton/DeepHSI)
Master: StoneFranklin/DeepHyperX#
To be used when running locally in a Python environment. It contains the timing metric.
Singularity: StoneFranklin/DeepHyperX#
The same as master branch, except with altered file paths for running in a singularity container.
Corresponding container: DeepHyperX-Timed.sif
Singularity-images: StoneFranklin/DeepHyperX#
The same as the singularity branch, with PNG image output added.
Corresponding container: DeepHyperX-Images.sif
Gateway-styled-output: StoneFranklin/DeepHyperX#
This branch outputs a text file with the results of the experiment. This is for use through the gateway.
Corresponding container: Gateway.sif
Thundersvm: StoneFranklin/DeepHyperX#
This branch contains a python file for testing thundersvm, as well as an altered definition file.
Corresponding container: cuda_test.sif
Aum-dataset: StoneFranklin/DeepHyperX#
Same as Gateway-styled-output, except with an altered custom_datasets.py file to allow the AUM dataset to be used.
Corresponding container: AUM.sif
ThreeLayer Classification: semihdinc/DeepHyperX#
This branch contains code from Dr. Dinc, for the threeLayer model based on the original DeepHyperX code.
DeepHyperX-aum-dataset_combined-threeLayer: 5MI7th3MI6/DeepHyperX-aum-dataset_combined-threeLayer#
This branch contains the working threeLayer classification model with AUM dataset. This combines the branch from AUM dataset with the threeLayer classification branch.
Corresponding container: aum-dataset.sif
Year: 2022#
DeepHyperX-aum-dataset_combined-threeLayer (updated): Fennrii/AUM-UNG-HSI-Repository#
This branch changed the threeLayer model to allow it to be run using different datasets, and also updated models.py to allow for larger datasets
Initial container: aum-dataset.sif:version1
Current container: aum-dataset.sif:version2
New container based on aum-dataset.sif:version2.#
Added parralellism threeLayer method based on numba as well as the ability to use the SVM_grid model in testing.
Current container: aum-dataset.sif:latest
Modified Jigsaw Code: Deepp0925/jigsaw#
Added new datasets, captures the time elapsed for each run
To run bash file:
sbatch ./multi_run.bash DATASET [Num of runs]
To run multiple datasets at once modify the Datasets array in assign_multi_run.bash with desired values and number of runs.
To run in interactive mode (only one run):
Activate the conda environment – conda activate con_jigsaw.env (if not exist run: conda create –name con_jigsaw_env –file conda_requisites.txt)
Python jigsaw.py DATASET
To add new dataset, modify the ‘get_target’ function and ‘readData’ function in jigsaw.py, add the parameters in config.ini