This code is associated with the paper from DeepFRET, a software for rapid and automated single molecule FRET data classification using deep learning et al., "Johannes Thomsen". eLife, 2020. http://doi.org/10.7554/eLife.60404

DeepFRET-Model

Repository for the model training.

Original Repo with ongoing development: here

To reproduce publication model:

1. Make sure your environment reflects requirement.txt
2. Run generate_data.py and input the number of traces (we used 250k initially, which ends up being ~150k after balancing)
3. Run train_model.py. Be sure to set exclude_alex_fret=False for the full model.