A repo for kindergarten curriculum learning in RNNs.
- dynamics: main python module
- process/rnn/: core code for training and simulating data from RNNs
- analysis: behavioral, state-space, and dynamics analsyis of RNNs
- utils: default configurations, RNN input specs, helper files
- vis: helper plotting functions
- demo: RNN training demos. for training and stimulating from an RNN model
- vis: Figure generation scripts for associated manuscript
-rat data: data from Mah et al. Zenodo repository, containing rat behavioral data: Mah, A., Schiereck, S., Bossio, V., & Constantinople, C. (2023). Distinct value computations support rapid sequential decisions (Version v1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.10031483
-RNN data: trained RNN model files and behavioral data, organized by CL type. to be published on Zenodo