Top, iedereen blij hoop ik

There is now a dataset card on @hf.co, but no data... huggingface.co/datasets/HPL...

Omg thank you. You saved the icon from being booted of my Home Screen for being incredibly out of place

Wel fijn dat je feature requests tegenwoordig ook in de krant kwijt kunt

You can find our models on Hugging Face (huggingface.co/NetherlandsF...) and the code on GitHub (github.com/NetherlandsF...)

First we trained a new base model from scratch called ARM64BERT. We then finetuned this base model for the binary code similarity task using the Sentence Transformers library.

Based on the jTrans x86-model by Wang et al that is suitable for the processors used in most personal computers, we've created a whole new tokenizer for the ARM64-code used in smartphones and embedded devices.

This automation saves a lot of time when reverse engineering binary code, which they do in order to get access to phones, to decrypt databases, or to simply get a better understanding of certain applications.

With this model our digital experts can turn a ARM64 assembly function into an embedding vector, use it as a query and find the most similar functions in a database of known functions.