SAFE
SAFE copied to clipboard
Can you provide a trained model for cross-architecture?
I want to compare two functions from different architecture, e.g., x86, ARM. I do not know how to train the model for cross-architecture. Can you provide a trained model for cross-architecture? Thank you.
I have trained this model with data in "AMD64ARMOpenSSL". The training command is as following:
python train.py -o ../out -e ../data/i2v/ -n ../data/AMD64ARMOpenSSL/AMD64ARMOpenSSL.db
Is it correct? I use the trained model in real-world cross-architecture binaries (arm and i586), but the result is poor. Is something wrong? Thank you.