ONNX model

1.0

Handling

Automatic release of the model

1,0

Handling

model (april release)

The model dataset is normalized to avoid "coin-flip" result side-effects. To get the full model, use the all.h5, rename it to original23.h5 to use it.

Models:

• original23.h5: Trained with 49 benign (randomly selected) and 49 malign binaries
• all.h5: Trained with 144 benign and 49 malign binaries.

mode (+unopt)

This model is trained with the original ralph binaries and the optimized cn ones

model (opt+unopt)

This model is trained with optimized and unoptimized malwares.

model

Here we place the model and the datasets.