interpret-benchmark
Benchmarking interpretation methods
Data can be found at http://mitra.stanford.edu/kundaje/eprakash/interpret-benchmark/final/simulation_pipeline/data/
DNN models trained with momma dragonn https://github.com/kundajelab/momma_dragonn
Pretrained models can be found at http://mitra.stanford.edu/kundaje/eprakash/interpret-benchmark/old/models/
DNN model config (for training with momma_dragonn) can be found at http://mitra.stanford.edu/kundaje/eprakash/interpret-benchmark/data/dnase_positives/momma_dragonn_config/deepsea_beluga/A549/sim_pos_and_neg/config/
The momma_dragonn package lives at: https://github.com/kundajelab/momma_dragonn
Simulated sequence models can be found at http://mitra.stanford.edu/kundaje/eprakash/interpret-benchmark/final/simulation_pipeline/models/
Interpretation scores can be found at http://mitra.stanford.edu/kundaje/eprakash/interpret-benchmark/final/simulation_pipeline/interpretation/