This project came up when i was following the advanced machine learning 2024 course at Sapienza university of Rome.
DeRuster is proudly powered by Compy, please refer to the documentation to apply changes.
The project's scripts need a folder named "data", inside this folder divide the datasets into two other folders: test and training. The training and validation sets are split automatically from the test folder inside two subsets, which the validation is TRAIN_SIZE/100*VALIDATION_SPLIT, validation split is set at 20%, it can be modified inside the dataset.py file.
This project makes use of python scripts, to run the model on a given binary please type in terminal:
run <binary_path>
To retrain the model type:
train
A dataset generator is provided, to generate functions and variables inside the data folder please type in terminal:
generate --size <n-programs>
To compile such dataset into the asm folder please type in terminal:
compile
This project came up when i was following the advanced machine learning 2024 course at Sapienza university of Rome.
DeRuster is proudly powered by Compy, please refer to the documentation to apply changes.
The project's scripts need a folder named "data", inside this folder divide the datasets into two other folders: test and training. The training and validation sets are split automatically from the test folder inside two subsets, which the validation is TRAIN_SIZE/100*VALIDATION_SPLIT, validation split is set at 20%, it can be modified inside the dataset.py file.
This project makes use of python scripts, to run the model on a given binary please type in terminal:
run <binary_path>
To retrain the model type:
train
A dataset generator is provided, to generate functions and variables inside the data folder please type in terminal:
generate --size <n-programs>
To compile such dataset into the asm folder please type in terminal:
compile