regLM is a toolkit for training hyenaDNA-based autoregressive language models on DNA sequences and generating novel regulatory elements.
To use regLM, first install HyenaDNA from GitHub following the instructions: https://github.com/HazyResearch/hyena-dna
git clone https://github.com/Genentech/regLM.git
cd regLM
pip install .
https://www.biorxiv.org/content/10.1101/2024.02.14.580373
Code used to perform the experiments in the regLM paper, along with trained model weights and synthetic sequences, are available at Zenodo.