Code for training and inferring the TAP model brain.
Dependencies: Python 3.5.6 or newer, PyTorch 1.5.1 or newer.
This folder contains the main python code:
rnnmodel.py: RNN model for the TAP braintapdynamics.py: functions for generating the TAP dynamicsparticlefilter.py: particle filter and Q function required for particle-EMnotebookutils.py: functions for loading a TAP brain, generating model parameters, etcutils.py: all other required utils
This folder contains the parameters for a few example TAP brains.
This folder contains particle-EM estimates for a few example TAP brains.
TrainTAPbrain.ipynb: example notebook that illustrates the training of a TAP brain modelInferTAPbrain.ipynb: example notebook that illustrates inferring inference in a TAP brain modelGreedyBackwardSearch.ipynb: example notebook that illustrates the greedy approach to find the best set of G and J parameters
InferTAPbrain.py: script for inferring inference in a TAP brain modelrunscripts.sh: shell script for launching multiple inference jobs in parallel